As Audio Large Language Models (ALLMs) emerge as powerful tools for speech processing, their safety implications demand urgent attention. While considerable research has explored textual and vision safety, audio’s distinct characteristics present significant challenges. This paper first investigates: Is ALLM vulnerable to backdoor attacks exploiting acoustic triggers? In response to this issue, we introduce Hidden in the Noise (HIN), a novel backdoor attack framework designed to exploit subtle, audio-specific features. HIN applies acoustic modifications to raw audio waveforms, such as alterations to temporal dynamics and strategic injection of spectrally tailored noise. These changes introduce consistent patterns that an ALLM’s acoustic feature encoder captures, embedding robust triggers within the audio stream. To evaluate ALLM robustness against audio-feature-based triggers, we develop the AudioSafe benchmark, assessing nine distinct risk types. Extensive experiments on AudioSafe and three established safety datasets reveal critical vulnerabilities in existing ALLMs: (I) audio features like environment noise and speech rate variations achieve over 90% average attack success rate, (II) ALLMs exhibit significant sensitivity differences across acoustic features, particularly showing minimal response to volume as a trigger, and (III) poisoned sample inclusion causes only marginal loss curve fluctuations, highlighting the attack’s stealth.

Recent progress in robotics and embodied AI is largely driven by Large Multimodal Models (LMMs). However, a key challenge remains underexplored: how can we advance LMMs to discover tasks that assist humans in open-future scenarios, where human intentions are highly concurrent and dynamic. In this work, we formalize the problem of Human-centric Open-future Task Discovery (HOTD), focusing particularly on identifying tasks that reduce human effort across plausible futures. To facilitate this study, we propose HOTD-Bench, which features over 2K real-world videos, a semi-automated annotation pipeline, and a simulation-based protocol tailored for open-set future evaluation. Additionally, we propose the Collaborative Multi-Agent Search Tree (CMAST) framework, which decomposes complex reasoning through a multi-agent system and structures the reasoning process through a scalable search tree module. In our experiments, CMAST achieves the best performance on the HOTD-Bench, significantly surpassing existing LMMs. It also integrates well with existing LMMs, consistently improving performance.

Liang Lin, Miao Yu, Moayad Aloqaily, Zhenhong Zhou, Kun Wang, Linsey Pang, Prakhar Mehrotra, Qingsong Wen. Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers). 2026.

Zengbin Wang, Feng Xiong, Liang Lin, Xuecai Hu, Yong Wang, Yanlin Wang, Man Zhang, Xiangxiang Chu. Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers). 2026.

Zhihao Zhan, Yuhao Chen, Jiaying Zhou, Qinhan Lyu, Hao Liu, Keze Wang, Liang Lin, Guangrun Wang. Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers). 2026.

Yuanhe Zhang, Jiayu Tian, Yibo Zhang, Shilinlu Yan, Liang Lin, Zhenhong Zhou, Li Sun, Sen Su. Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers). 2026.

Standard Chain-of-Thought (CoT) prompting empowers Large Language Models (LLMs) with reasoning capabilities, yet its reliance on linear natural language is inherently insufficient for effective world modeling in embodied tasks. While text offers flexibility, it fails to explicitly represent the state-space, object hierarchies, and causal dependencies required for robust robotic planning. To address these limitations, we propose Object-Oriented World Modeling (OOWM), a novel framework that structures embodied reasoning through the lens of software engineering formalisms. We redefine the world model not as a latent vector space, but as an explicit symbolic tuple \(\mathcal {W} = \langle \mathcal {S}, \mathcal {T} \rangle\) : a State Abstraction (\(G_\text{state}\)) instantiating the environmental state \(\mathcal {S}\), coupled with a Control Policy (\(G_\text{control}\)) representing the transition logic \(\mathcal {T}: S \times A \rightarrow S^{\prime }\). OOWM leverages the Unified Modeling Language (UML) to materialize this definition: it employs Class Diagrams to ground visual perception into rigorous object hierarchies, and Activity Diagrams to operationalize planning into executable control flows. Furthermore, we introduce a three-stage training pipeline combining Supervised Fine-Tuning (SFT) with Group Relative Policy Optimization (GRPO). Crucially, this method utilizes outcome-based rewards from the final plan to implicitly optimize the underlying object-oriented reasoning structure, enabling effective learning even with sparse annotations. Extensive evaluations on the MRoom-30k benchmark demonstrate that OOWM significantly outperforms unstructured textual baselines in planning coherence, execution success, and structural fidelity, establishing a new paradigm for structured embodied reasoning.

The scarcity of large-scale robotic data has motivated the repurposing of foundation models from other modalities for policy learning. In this work, we introduce PhysGen (Learning Physics from Pretrained Video Generation Models), a scalable continuous and sequential world interaction framework that leverages autoregressive video generation to solve robotic manipulation tasks. By treating the pretrained video model as a proxy for a physics simulator, PhysGen models the dynamic interplay between the external environment and robot actions. We introduce a multimodal continuous representation that unifies video and action into shared physical tokens, bridging the gap between discrete video generation and continuous robotic control. This approach enables the seamless transfer of implicit physical knowledge—such as object permanence and dynamics—from video pretraining to downstream manipulation.To ensure efficient convergence, we incorporate causal masking, inverse kinematics, Lookahead Multi-Token Prediction (L-MTP), and key-value (KV) caching. Experimental results on the Libero and ManiSkill benchmarks demonstrate that PhysGen consistently outperforms robust baselines, surpassing OpenVLA and WorldVLA by margins of 13.8% and 8.8%, respectively. Notably, in real-world scenarios, PhysGen matches the performance of large-scale action-pretrained models like π0 without requiring prior action-specific pretraining, demonstrating superior capability in physically complex tasks such as grasping transparent objects. These findings validate the potential of extracting physical intuition from pretrained video generators to facilitate generalizable robotic manipulation.

Video-text retrieval is a fundamental task in multi-modal learning, aiming to accurately retrieve videos that match given textual descriptions. While recent contrastive methods have made significant progress by embedding videos and texts into a joint space, they often suffer from semantic over-clustering—a phenomenon where semantically distinct videos are mapped to overly similar embeddings due to dominant but uninformative visual patterns (e.g., recurring backgrounds or common objects). This effect becomes particularly problematic under short or ambiguous queries, where it suppresses fine-grained semantics and degrades retrieval precision. To address this, we propose Similarity-aware Probabilistic Embeddings Modeling (SPEM), a novel framework that refines video representations by modeling them as adaptive probability distributions rather than static vectors. SPEM incorporates cross-modal attention to high-light text-relevant visual content and suppress irrelevant patterns, and leverages multi-level similarity features to dynamically adjust the embedding variance, thereby preserving subtle but critical semantic cues. To further improve alignment, we employ a Semantic-Distribution Contrastive Loss to optimize the alignment structure in the probabilistic space, encouraging more discriminative separation across hard negatives. Extensive experiments on four widely-used video-text retrieval benchmarks—MSRVTT, DiDeMo, VATEX, and MSVD—demonstrate that SPEM consistently outperforms strong CLIP-based baselines.

Speech-preserving facial expression manipulation (SPFEM) aims to modify facial emotions while meticulously maintaining the mouth animation associated with spoken content. Current works depend on inaccessible paired training samples for the person, where two aligned frames exhibit the same speech content yet differ in emotional expression, limiting the SPFEM applications in real-world scenarios. In this work, we discover that speakers who convey the same content with different emotions exhibit highly correlated local facial animations in both spatial and temporal spaces, providing valuable supervision for SPFEM. To capitalize on this insight, we propose a novel spatial-temporal coherent correlation learning (STCCL) algorithm, which models the aforementioned correlations as explicit metrics and integrates the metrics to supervise manipulating facial expression and meanwhile better preserving the facial animation of spoken content. To this end, it first learns a spatial coherent correlation metric, ensuring that the visual correlations of adjacent local regions within an image linked to a specific emotion closely resemble those of corresponding regions in an image linked to a different emotion. Simultaneously, it develops a temporal coherent correlation metric, ensuring that the visual correlations of specific regions across adjacent image frames associated with one emotion are similar to those in the corresponding regions of frames associated with another emotion. Recognizing that visual correlations are not uniform across all regions, we have also crafted a correlation-aware adaptive strategy that prioritizes regions that present greater challenges. During SPFEM model training, we construct the spatial-temporal coherent correlation metric between corresponding local regions of the input and output image frames as additional loss to supervise the generation process. We conduct extensive experiments on various datasets, and the results demonstrate the effectiveness of the proposed STCCL algorithm.

Image fusion seeks to seamlessly integrate foreground objects with background scenes, producing realistic and harmonious fused images. While existing methods often insert objects directly, adaptive and interactive fusion-requiring contextual adaptation and foreground-background interplay-remains a challenging yet critical task. To address this, we first propose a pipeline for generating high-quality fusion data. By combining iterative in-context learning with existing tools, we curate a diverse cross-scene dataset supporting three core tasks: object integration, replacement, and attribute-referenced editing. Leveraging this, we introduce DreamFuse, a unified diffusion-based approach that jointly optimizes these capabilities. DreamFuse exploits the Diffusion Transformer (DiT) architecture, using its attention mechanism to extract and align foreground-background features for coherent fusion. For flexible control, we incorporate a Positional Affine mechanism, enabling precise spatial and scale adjustments while supporting diverse text-driven fusion. Furthermore, we employ Localized Direct Preference Optimization (L-DPO), refining the model via human feedback to enhance harmony and consistency. Extensive experimental results demonstrate DreamFuse's superiority over state-of-the-art approaches across multiple metrics.

Recent studies have revealed the potential of training open-source Large Language Models (LLMs) to unleash LLMs' reasoning ability for enhancing vision-language navigation (VLN) performance, and simultaneously mitigate the domain gap between LLMs' training corpus and the VLN task. However, these approaches predominantly adopt straightforward input-output mapping paradigms, causing the mapping learning difficult and the navigational decisions unexplainable. Chain-of-Thought (CoT) training is a promising way to improve both navigational decision accuracy and interpretability, while the complexity of the navigation task makes the perfect CoT labels unavailable and may lead to overfitting through pure CoT supervised fine-tuning. To address these issues, we propose EvolveNav, a novel sElf-improving embodied reasoning paradigm that realizes adaptable and generalizable navigational reasoning for boosting LLM-based vision-language Navigation. Specifically, EvolveNav involves a two-stage training process: (1) Formalized CoT Supervised Fine-Tuning, where we train the model with curated formalized CoT labels to first activate the model's navigational reasoning capabilities, and simultaneously increase the reasoning speed; (2) Self-Reflective Post-Training, where the model is iteratively trained with its own reasoning outputs as self-enriched CoT labels to enhance the supervision diversity. A self-reflective auxiliary task is also designed to encourage the model to learn correct reasoning patterns by contrasting with wrong ones. Experimental results under both task-specific and cross-task training paradigms demonstrate the consistent superiority of EvolveNav over previous LLM-based VLN approaches on various popular benchmarks, including R2R, REVERIE, CVDN, and SOON. EvolveNav open avenues for exploring effective self-improving reasoning paradigms, enabling building agents capable of self-evolving for promoting LLM-based embodied AI research.

Synthesizing human motion from textual descriptions is essential for immersive digital applications, yet existing methods face a persistent trade-off between semantic fidelity and physical realism. Large language model (LLM)-based approaches can interpret diverse open-vocabulary instructions and compose high-level action plans, but they often generate motions that violate physical constraints. Physics-aware models improve realism through simulation or control, but they struggle with semantic complexity, fine-grained instructions, and novel concepts. To address this gap, we propose In-Context Model Predictive Generation (ICMPG), a framework that integrates language-model planning with inference-time physical feedback. ICMPG reformulates motion synthesis as a Model Predictive Control (MPC)-like process with two modules. The Context-Aware Motion Generation (CAMG) module uses an LLM as a planner to decompose textual commands and generate candidate motion sequences from motion tokens. The Model Predictive Generation (MPG) module evaluates these candidates through physical simulation and semantic alignment, estimates a composite reward, and selects the best sequence to guide subsequent generation steps. Unlike open-loop generation, this closed-loop refinement enables ICMPG to adapt motions to both the input semantics and the simulated physical environment without task-specific policy retraining. Extensive experiments across standard and zero-shot open-vocabulary settings show that ICMPG generalizes robustly to diverse commands and produces motions that are more physically plausible and semantically faithful than representative baselines on the evaluated benchmarks. The framework bridges semantic interpretation and physical simulation while remaining flexible enough to incorporate different LLM backbones, enabling more versatile and controllable text-driven motion synthesis.

Pre-trained Vision-Language Models (VLMs) have demonstrated strong zero-shot generalization capabilities. Despite their effectiveness on various downstream tasks, they remain vulnerable to adversarial samples. Existing methods fine-tune VLMs to improve their robust performance by performing adversarial training on a certain dataset. However, this can lead to model overfitting and is not a true zero-shot scenario. In this paper, we propose a truly zero-shot and training-free approach that can improve the zero-shot adversarial robustness of VLMs on the evaluated benchmarks. Specifically, we first discover that simply adding Gaussian noise can enhance the VLM's zero-shot robustness. Then, we treat the adversarial examples with added Gaussian noise as anchors and strive to find a path in the embedding space that leads from the adversarial examples to the cleaner samples. Furthermore, to avoid the overfitting issue caused by fixed hyperparameters, we propose an adaptive parameter adjustment method based on the distance between the anchors and adversarial samples in the embedding space. We largely preserve the original VLMs' zero-shot generalization abilities in a truly zero-shot and training-free manner on the evaluated benchmarks compared to previous methods. Extensive experiments on 16 datasets demonstrate that our method can achieve stronger zero-shot robust performance, improving the top-1 robust accuracy by an average of 10.83%.

A few recent works attempt to train an adversarially robust Unsupervised Domain Adaptation (UDA) model, transferring the robustness from a robust source model or other robust pre-trained models to an unlabeled target domain. However, it is usually impractical to assume the availability of robust source models or robust pre-training, and meanwhile, source data are not always accessible or efficient for adaptation training in many real-world scenarios. In this paper, we dive into a more practical and challenging problem of robust source-free domain adaptation: can we train a robust model on an unlabeled target domain given only a non-robust source model (without source data)? Empirically, we find that applying adversarial training (AT) to the self-supervised adaptation process leads to severe model degradation, as it tends to amplify the inevitable errors of UDA models. To tackle this issue, we propose a novel approach called Source-Free Alternating Optimization (SFAO), which employs a non-robust target model to provide better guidance for the AT of the desired robust target model. The two models are trained in an alternating manner to minimize the discrepancy between the clean source domain and the adversarial target domain. Moreover, we propose Softly-Constrained Adversarial Training (SCAT) to further mitigate the adverse effects of incorrect pseudo-labels in AT. Extensive experimental results demonstrate that the proposed method significantly improves the model performance on both clean and adversarial data. Source code is available at: https://github.com/Coxy7/robust-SFDA.

Audio-driven talking face video generation has attracted increasing attention due to its huge industrial potential. Some previous methods focus on learning a direct mapping from audio to visual content. Despite progress, they often struggle with the ambiguity of the mapping process, leading to flawed results. An alternative strategy involves facial structural representations (e.g., facial landmarks) as intermediaries. This multi-stage approach better preserves the appearance details but suffers from error accumulation due to the independent optimization of different stages. Moreover, most previous methods rely on generative adversarial networks, prone to training instability and mode collapse. To address these challenges, our study proposes a novel landmark-based diffusion model for talking face generation, which leverages facial landmarks as intermediate representations while enabling end-to-end optimization. Specifically, we first establish the less ambiguous mapping from audio to landmark motion of lip and jaw. Then, we introduce an innovative conditioning module called TalkFormer to align the synthesized motion with the motion represented by landmarks via differentiable cross-attention, which enables end-to-end optimization for improved lip synchronization. Besides, TalkFormer employs implicit feature warping to align the reference image features with the target motion for preserving more appearance details. Extensive experiments demonstrate that our approach can synthesize high-fidelity and lip-synced talking face videos, preserving more subject appearance details from the reference image.

Recent methods to enhance Vision-Language Models (VLMs) for Visual Question Answering (VQA) have focused on strengthening their inference capabilities, enabling them to tackle VQA tasks independently rather than merely as aids to Large Language Models (LLMs). However, these approaches often ignore the rich commonsense knowledge inside the given VQA image sampled from the real world, limiting the full potential of VLMs. Inspired by the human top-down reasoning process, i.e., systematically exploring relevant issues to derive a comprehensive answer, this work introduces a novel, explainable multi-agent collaboration framework by leveraging the expansive knowledge of LLMs to enhance the capabilities of VLMs themselves. Our framework comprises three agents, i.e.,Responder,Seeker, andIntegrator, to collaboratively answer the given VQA question by seeking its relevant issues and generating the final answer in such a top-down reasoning process. The VLM-basedResponderagent generates the answer candidates for the question and responds to other relevant issues. TheSeekeragent, primarily based on LLM, identifies relevant issues related to the question to inform theResponderagent and constructs a Multi-View Knowledge Base (MVKB) for the given visual scene by leveraging the build-in world knowledge of LLM. TheIntegratoragent combines knowledge from theSeekeragent and theResponderagent to produce the final VQA answer. Extensive and comprehensive evaluations on diverse VQA datasets with a variety of VLMs demonstrate the superior performance and interpretability of our framework over the baseline method, e.g., 5.7% improvement on VQA-RAD and 5.2% on Winoground in the zero-shot setting without extra training cost.

Deductive reasoning is a crucial logical capability that assists us in solving complex problems based on existing knowledge. Although augmented by Chain-of-Thought prompts, Large Language Models (LLMs) might not follow the correct reasoning paths. Enhancing the deductive reasoning abilities of LLMs, and leveraging their extensive built-in knowledge for various reasoning tasks, remains an open question. Attempting to mimic the human deductive reasoning paradigm, we propose a multi-stage Syllogistic-Reasoning Framework of Thought (SR-FoT) that enables LLMs to perform syllogistic deductive reasoning to handle complex knowledge-based reasoning tasks. Our SR-FoT begins by interpreting the question and then uses the interpretation and the original question to propose a suitable major premise. It proceeds by generating and answering minor premise questions in two stages to match the minor premises. Finally, it guides LLMs to use the previously generated major and minor premises to perform syllogistic deductive reasoning to derive the answer to the original question. Extensive and thorough experiments on knowledge-based reasoning tasks have demonstrated the effectiveness and advantages of our SR-FoT.

This paper focuses on the newly emerged research topic, i.e., whether the complex decision-making logic of a DNN can be mathematically summarized into a few simple logics. Beyond the explanation of a static DNN, in this paper, we hope to show that the seemingly complex learning dynamics of a DNN can be faithfully represented as the change of a few primitive interaction patterns encoded by the DNN. Therefore, we redefine the interaction of principal feature components in intermediate-layer features, which enables us to concisely summarize the highly complex dynamics of interactions throughout the learning of the DNN. The mathematical faithfulness of the new interaction is experimentally verified. From the perspective of learning efficiency, we find that the interactions naturally belong to five groups (reliable, withdrawn, forgotten, betraying, and fluctuating interactions), each representing a distinct type of dynamics of an interaction being learned and/or being forgotten. This provides deep insights into the learning process of a DNN.

Long Context Understanding (LCU) is a critical area for exploration in current large language models (LLMs).However, due to the inherently lengthy nature of long-text data, existing LCU benchmarks for LLMs often result in prohibitively high evaluation costs, like testing time and inference expenses.Through extensive experimentation, we discover that existing LCU benchmarks exhibit significant redundancy, which means the inefficiency in evaluation.In this paper, we propose a concise data compression method tailored for longtext data with sparse information characteristics.By pruning the well-known LCU benchmark LongBench, we create MiniLongBench.This benchmark includes only 237 test samples across six major task categories and 21 distinct tasks.Through empirical analysis of over 60 LLMs, MiniLongBench achieves an average evaluation cost reduced to only 4.5% of the original while maintaining an average rank correlation coefficient of 0.97 with Long-Bench results.Therefore, our MiniLongBench, as a low-cost benchmark, holds great potential to substantially drive future research into the LCU capabilities of LLMs.See Github for our code, data and tutorial.

We focus on the problem of fusing two or more heterogeneous large language models (LLMs) to leverage their complementary strengths.One of the challenges of model fusion is high computational load, specifically in fine-tuning or aligning vocabularies.To address this, we propose Cool-Fusion, a simple yet effective approach that fuses the knowledge of source LLMs, which does not require training.Unlike ensemble methods, Cool-Fusion is applicable to any set of source LLMs that have different vocabularies.To overcome the vocabulary discrepancies among LLMs, we ensemble LLMs on text level, allowing them to rerank the generated texts by each other with different granularities.Extensive experiments have been conducted across a variety of benchmark datasets.On GSM8K, Cool-Fusion increases accuracy from three strong source LLMs by a significant margin of 17.4%.

Contrastive Language-Image Pre-training (CLIP) models as a milestone of modern multimodal intelligence, its gener-alization mechanism grasped massive research interests in the community. While existing studies limited in the scope of pre-training knowledge, hardly underpinned its generalization to countless open-world concepts absent from the pre-training regime. This paper dives into such Out-of-Pre-training (OOP) generalization problem from a holistic perspective. We propose LAION-Beyond benchmark to isolate the evaluation of OOP concepts from pre-training knowledge, with regards to OpenCLIP and its reproducible variants derived from LAION datasets. Empirical analysis evidences that despite image features of OOP concepts born with significant category margins, their zero-shot transfer significantly fails due to the poor image-text alignment. To this, we elaborate the "name-tuning" methodology with its theoretical merits in terms of OOP generalization, then propose few-shot name learning (FSNL) and zero-shot name learning (ZSNL) algorithms to achieve OOP generalization in a data-efficient manner. LAION-Beyond dataset and codes: http://m-huangx.github.io/laion_beyond/.

Video question grounding (VideoQG) requires models to answer the questions and simultaneously infer the relevant video segments to support the answers. However, existing VideoQG methods usually suffer from spurious cross-modal correlations, leading to a failure to identify the dominant visual scenes that align with the intended question. Moreover, vision-language models exhibit unfaithful generalization performance and lack robustness on challenging downstream tasks such as VideoQG. In this work, we propose a novel VideoQG framework named Cross-modal Causal Relation Alignment (CRA), to eliminate spurious correlations and improve the causal consistency between question-answering and video temporal grounding. Our CRA involves three essential components: i) Gaussian Smoothing Grounding (GSG) module for estimating the time interval via cross-modal attention, which is de-noised by an adaptive Gaussian filter, ii) Cross-Modal Alignment (CMA) enhances the performance of weakly supervised VideoQG by leveraging bidirectional contrastive learning between estimated video segments and QA features, iii) Explicit Causal Intervention (ECI) module for multimodal deconfounding, which involves front-door intervention for vision and backdoor intervention for language. Extensive experiments on two VideoQG datasets demonstrate the superiority of our CRA in discovering visually grounded content and achieving robust question reasoning. Codes are available at https://github.com/WissingChen/CRA-GQA.

Virtual Try-On (VTON) is a transformative technology in e-commerce and fashion design, enabling realistic digital visualization of clothing on individuals. In this work, we propose VTON 360, a novel 3D VTON method that addresses the open challenge of achieving high-fidelity VTON that supports any-view rendering. Specifically, we leverage the equivalence between a 3D model and its rendered multi-view 2D images, and reformulate 3D VTON as an extension of 2D VTON that ensures 3D consistent results across multiple views. To achieve this, we extend 2D VTON models to include multi-view garments and clothing-agnostic human body images as input, and propose several novel techniques to enhance them, including: i) a pseudo-3D pose representation using normal maps derived from the SMPL-X 3D human model, ii) a multi-view spatial attention mechanism that models the correlations between features from different viewing angles, and iii) a multi-view CLIP embedding that enhances the garment CLIP features used in 2D VTON with camera information. Extensive experiments on large-scale real datasets and clothing images from e-commerce platforms demonstrate the effectiveness of our approach. Project page: https://scnuhealthy.github.io/VTON360.

3D Question Answering (3D QA) requires the model to comprehensively understand its situated 3D scene described by the text, then reason about its surrounding environment and answer a question under that situation. However, existing methods usually rely on global scene perception from pure 3D point clouds and overlook the importance of rich local texture details from multi-view images. Moreover, due to the inherent noise in camera poses and complex occlusions, there exists significant feature degradation and reduced feature robustness problems when aligning 3D point cloud with multi-view images. In this paper, we propose a Dual-vision Scene Perception Network (DSPNet), to comprehensively integrate multi-view and point cloud features to improve robustness in 3D QA. Our Text-guided Multi-view Fusion (TGMF) module prioritizes image views that closely match the semantic content of the text. To adaptively fuse back-projected multi-view images with point cloud features, we design the Adaptive Dual-vision Perception (ADVP) module, enhancing 3D scene comprehension. Additionally, our Multimodal Context-guided Reasoning (MCGR) module facilitates robust reasoning by integrating contextual information across visual and linguistic modalities. Experimental results on SQA3D and ScanQA datasets demonstrate the superiority of our DSPNet. Codes will be available at https://github.com/LZ-CH/DSPNet.

Existing Vision-Language Navigation (VLN) methods primarily focus on single-stage navigation, limiting their effectiveness in multi-stage and long-horizon tasks within complex and dynamic environments. To address these limitations, we propose a novel VLN task, named Long-Horizon Vision-Language Navigation (LH-VLN), which emphasizes long-term planning and decision consistency across consecutive subtasks. Furthermore, to support LH-VLN, we develop an automated data generation platform NavGen, which constructs datasets with complex task structures and improves data utility through a bidirectional, multi-granularity generation approach. To accurately evaluate complex tasks, we construct the Long-Horizon Planning and Reasoning in VLN (LHPR-VLN) benchmark consisting of 3,260 tasks with an average of 150 task steps, serving as the first dataset specifically designed for the long-horizon vision-language navigation task. Furthermore, we propose Independent Success Rate (ISR), Conditional Success Rate (CSR), and CSR weight by Ground Truth (CGT) metrics, to provide fine-grained assessments of task completion. To improve model adaptability in complex tasks, we propose a novel Multi-Granularity Dynamic Memory (MGDM) module that integrates short-term memory blurring with long-term memory retrieval to enable flexible navigation in dynamic environments. Our platform, benchmark and method supply LH-VLN with a robust data generation pipeline, comprehensive model evaluation dataset, reasonable metrics, and a novel VLN model, establishing a foundational framework for advancing LH-VLN.

Text-driven avatar generation has gained significant attention owing to its convenience. However, existing methods typically model the human body with all garments as a single 3D model, limiting its usability, such as clothing replacement, and reducing user control over the generation process. To overcome the limitations above, we propose DAGSM, a novel pipeline that generates disentangled human bodies and garments from the given text prompts. Specifically, we model each part (e.g., body, upper/lower clothes) of the clothed human as one GS-enhanced mesh (GSM), which is a traditional mesh attached with 2D Gaussians to better handle complicated textures (e.g., woolen, translucent clothes) and produce realistic cloth animations. During the generation, we first create the unclothed body, followed by a sequence of individual cloth generation based on the body, where we introduce a semantic-based algorithm to achieve better human-cloth and garment-garment separation. To improve texture quality, we propose a view-consistent texture refinement module, including a cross-view attention mechanism for texture style consistency and an incident-angle-weighted denoising (IAW-DE) strategy to update the appearance. Extensive experiments have demonstrated that DAGSM generates high-quality disentangled avatars, supports clothing replacement and realistic animation, and outperforms the baselines in visual quality.

Recent video editing advancements rely on accurate pose sequences to animate human actors. However, these efforts are not suitable for cross-species animation due to pose misalignment between species (for example, the poses of a cat differ greatly from that of a pig due to their distinct body structures). In this paper, we present Anima2, a zero-shot diffusion-based video generator to address this issue, aiming to accurately ANIMAte ANIMAls while preserving the background. The key technique involves two-fold subject alignment. First, we improve appearance feature extraction by integrating a Laplacian detail booster and a prompt-tuning identity extractor. They capture essential appearance information, including identity and fine details. Second, we align shape features and address conflicts from differing animals by introducing a scale-information remover and an adaptive rescaling module. They both enhance subject alignment for accurate cross-species animation. Additionally, we introduce two high-quality animal video datasets with diverse species to benchmark cross-species animation. Trained on these extensive datasets, our model directly generates videos with accurate movements, consistent appearances, and high-fidelity frames, eliminating the need for test-time training. Extensive experiments demonstrate our method’s superiority in cross-species animation, showcasing robust adaptability and generality.

Contemporary diffusion models built upon U-Net or Diffusion Transformer (DiT) architectures have revolutionized image generation through transformer-based attention mechanisms. The prevailing paradigm has commonly employed self-attention with quadratic computational complexity to handle global spatial relationships in complex images, thereby synthesizing high-fidelity images with coherent visual semantics.Contrary to conventional wisdom, our systematic layer-wise analysis reveals an interesting discrepancy: self-attention in pre-trained diffusion models predominantly exhibits localized attention patterns, closely resembling convolutional inductive biases. This suggests that global interactions in self-attention may be less critical than commonly assumed.Driven by this, we propose \(Δ\)ConvFusion to replace conventional self-attention modules with Pyramid Convolution Blocks (\(Δ\)ConvBlocks).By distilling attention patterns into localized convolutional operations while keeping other components frozen, \(Δ\)ConvFusion achieves performance comparable to transformer-based counterparts while reducing computational cost by 6929$\times$ and surpassing LinFusion by 5.42$\times$ in efficiency--all without compromising generative fidelity.

Image fusion seeks to seamlessly integrate foreground objects with background scenes, producing realistic and harmonious fused images. Unlike existing methods that directly insert objects into the background, adaptive and interactive fusion remains a challenging yet appealing task. It requires the foreground to adjust or interact with the background context, enabling more coherent integration. To address this, we propose an iterative human-in-the-loop data generation pipeline, which leverages limited initial data with diverse textual prompts to generate fusion datasets across various scenarios and interactions, including placement, holding, wearing, and style transfer. Building on this, we introduce DreamFuse, a novel approach based on the Diffusion Transformer (DiT) model, to generate consistent and harmonious fused images with both foreground and background information. DreamFuse employs a Positional Affine mechanism to inject the size and position of the foreground into the background, enabling effective foreground-background interaction through shared attention. Furthermore, we apply Localized Direct Preference Optimization guided by human feedback to refine DreamFuse, enhancing background consistency and foreground harmony. DreamFuse achieves harmonious fusion while generalizing to text-driven attribute editing of the fused results. Experimental results demonstrate that our method outperforms state-of-the-art approaches across multiple metrics.

Embodied Question Answering (EQA) is a challenging task in embodied intelligence that requires agents to dynamically explore 3D environments, actively gather visual information, and perform multi-step reasoning to answer questions. However, current EQA approaches suffer from critical limitations in exploration efficiency, dataset design, and evaluation metrics. Moreover, existing datasets often introduce biases or prior knowledge, leading to disembodied reasoning, while frontier-based exploration strategies struggle in cluttered environments and fail to ensure fine-grained exploration of task-relevant areas. To address these challenges, we construct the EXPloration-awaRe Embodied queStion anSwering Benchmark (EXPRESS-Bench), the largest dataset designed specifically to evaluate both exploration and reasoning capabilities. EXPRESS-Bench consists of 777 exploration trajectories and 2,044 question-trajectory pairs. To improve exploration efficiency, we propose Fine-EQA, a hybrid exploration model that integrates frontier-based and goal-oriented navigation to guide agents toward task-relevant regions more effectively. Additionally, we introduce a novel evaluation metric, Exploration-Answer Consistency (EAC), which ensures faithful assessment by measuring the alignment between answer grounding and exploration reliability. Extensive experimental comparisons with state-of-the-art EQA models demonstrate the effectiveness of our EXPRESS-Bench in advancing embodied exploration and question reasoning.

Temporal sentence grounding aims to identify exact moments in a video that correspond to a given textual query, typically addressed with detection transformer (DETR) solutions. However, we find that typical strategies designed to enhance DETR do not improve, and may even degrade, its performance in this task. We systematically analyze and identify the root causes of this abnormal behavior: (1) conflicts between queries from similar target moments and (2) internal query conflicts due to the tension between global semantics and local localization. Building on these insights, we propose a simple yet powerful baseline, Sim-DETR, which extends the standard DETR with two minor modifications in the decoder layers: (1) constraining self-attention between queries based on their semantic and positional overlap and (2) adding query-to-frame alignment to bridge the global and local contexts. Experiments demonstrate that Sim-DETR unlocks the full potential of DETR for temporal sentence grounding, offering a strong baseline for future research.

Video Multimodal Large Language Models~(Video-MLLM) have achieved remarkable advancements in video understanding tasks. However, constrained by the context length limitation in the underlying LLMs, existing Video-MLLMs typically exhibit suboptimal performance on long video scenarios. To understand extended input frames, common solutions span token compression and streaming inference techniques, which sacrifice feature granularity or inference efficiency. Differently, to efficiently achieve comprehensive understanding of longer frame inputs, we draw ideas from MoE and propose a training-free approach \textbf{Free-MoRef}, which instantly multiplexes the context perception capabilities of Video-MLLMs within one inference pass. Specifically, Free-MoRef reconstructs the vision tokens into several short sequences as multi-references. Subsequently, we introduce MoRef-attention, which gathers clues from the multi-reference chunks in parallel to summarize unified query activations. After the shadow layers in LLMs, a reference fusion step is derived to compose a final mixed reasoning sequence with key tokens from parallel chunks, which compensates the cross-reference vision interactions that are neglected in MoRef-attention. By splitting and fusing the long vision token sequences, Free-MoRef achieves improved performance under much lower computing costs in reasoning multiplexed context length, demonstrating strong efficiency and effectiveness. Experiments on VideoMME, MLVU, LongVideoBench show that Free-MoRef achieves full perception of 2$\times$ to 8$\times$ longer input frames without compression on a single A100 GPU while keeping instant responses, thereby bringing significant performance gains, even surpassing dedicatedly trained long-video-MLLMs. Codes are available at https://github.com/wkfdb/Free-MoRef

Diffusion-Based Purification (DBP) has emerged as an effective defense mechanism against adversarial attacks. The success of DBP is often attributed to the forward diffusion process, which reduces the distribution gap between clean and adversarial images by adding Gaussian noise. Although this explanation is theoretically grounded, the precise contribution of this process to robustness remains unclear. In this paper, through a systematic investigation, we propose that the intrinsic stochasticity in the DBP procedure is the primary factor driving robustness. To explore this hypothesis, we introduce a novel Deterministic White-Box (DW-box) evaluation protocol to assess robustness in the absence of stochasticity, and analyze attack trajectories and loss landscapes. Our results suggest that DBP models primarily leverage stochasticity to evade effective attack directions, and that their ability to purify adversarial perturbations can be weak. To further enhance the robustness of DBP models, we propose Adversarial Denoising Diffusion Training (ADDT), which incorporates classifier-guided adversarial perturbations into diffusion training, thereby strengthening the models' ability to purify adversarial perturbations. Additionally, we propose Rank-Based Gaussian Mapping (RBGM) to improve the compatibility of perturbations with diffusion models. Experimental results validate the effectiveness of ADDT. In conclusion, our study suggests that future research on DBP can benefit from the perspective of decoupling stochasticity-based and purification-based robustness.

Despite advancing language model (LM) alignment, direct preference optimization (DPO) falls short in LM reasoning with the free lunch from reinforcement learning (RL). As the breakthrough, this work proposes a new RL-free preference optimization method aiming to achieve DPO along with learning another LM, whose response generation policy holds the asymptotic equivalence with AlphaZero-like search, the apex of algorithms for complex reasoning missions like chess Go. While circumventing explicit value and reward modeling, the neural implicit tree search executed by the extra LM remains seeking to equip DPO with reasoning procedure technically akin to AlphaZero. Our experiments demonstrate that our methodology outperforms both regular DPO variants in human preference alignment, and MCTS-based LMs in mathematical reasoning and planning tasks.

The remarkable evolution of generative models has enabled the generation of high-quality, visually attractive images, often perceptually indistinguishable from real photographs to human eyes. This has spurred significant attention on AI-generated image (AIGI) detection. Intuitively, higher image quality should increase detection difficulty. However, our systematic study on cutting-edge text-to-image generators reveals a counterintuitive finding: AIGIs with higher quality scores, as assessed by human preference models, tend to be more easily detected by existing models. To investigate this, we examine how the text prompts for generation and image characteristics influence both quality scores and detector accuracy. We observe that images from short prompts tend to achieve higher preference scores while being easier to detect. Furthermore, through clustering and regression analyses, we verify that image characteristics like saturation, contrast, and texture richness collectively impact both image quality and detector accuracy. Finally, we demonstrate that the performance of off-the-shelf detectors can be enhanced across diverse generators and datasets by selecting input patches based on the predicted scores of our regression models, thus substantiating the broader applicability of our findings. Code and data are available at https://github.com/Coxy7/AIGI-Detection-Quality-Paradox.

Alpha mining, a critical component in quantitative investment, focuses on discovering predictive signals for future asset returns in increasingly complex financial markets. However, the pervasive issue of alpha decay-where factors lose their predictive power over time-poses a significant challenge for alpha mining. Traditional methods such as genetic programming are prone to rapid alpha decay, primarily due to their susceptibility to overfitting. At the same time, approaches driven by Large Language Models (LLMs), despite their promise, often fail to impose regularization against factor homogenization-resulting in crowded signals and accelerated decay. To address this challenge, we propose AlphaAgent, an autonomous framework that effectively integrates LLM-driven agents with ad hoc regularization for mining decay-resistant alpha factors. AlphaAgent employs three key mechanisms: (i) originality enforcement through a similarity measure based on abstract syntax trees (ASTs) against existing alphas(ii) hypothesis-factor alignment via LLM-evaluated semantic consistency between market hypotheses and generated factors, and (iii) complexity control via AST-based structural constraints, preventing over-engineered constructions that are prone to overfitting. These mechanisms collectively guide the alpha generation process to balance originality, financial rationale, and adaptability to evolving market conditions, mitigating the risk of alpha decay. Extensive evaluations show that AlphaAgent outperforms traditional and LLM-based methods in mitigating alpha decay across bull and bear markets, consistently delivering significant alpha in Chinese CSI 500 and U.S. S&P 500 markets over the past four years. Notably, AlphaAgent showcases remarkable resistance to alpha decay, elevating the potential for yielding powerful factors.

Open-Vocabulary Multi-Label Recognition (OV-MLR) aims to identify multiple seen and unseen object categories within an image, requiring both precise intra-class localization to pinpoint objects and effective inter-class reasoning to model complex category dependencies. While Vision-Language Pre-training (VLP) models offer a strong open-vocabulary foundation, they often struggle with fine-grained localization under weak supervision and typically fail to explicitly leverage structured relational knowledge beyond basic semantics, limiting performance especially for unseen classes. To overcome these limitations, we propose the Dual Adaptive Refinement Transfer (DART) framework. DART enhances a frozen VLP backbone via two synergistic adaptive modules. For intra-class refinement, an Adaptive Refinement Module (ARM) refines patch features adaptively, coupled with a novel Weakly Supervised Patch Selecting (WPS) loss that enables discriminative localization using only image-level labels. Concurrently, for inter-class transfer, an Adaptive Transfer Module (ATM) leverages a Class Relationship Graph (CRG), constructed using structured knowledge mined from a Large Language Model (LLM), and employs graph attention network to adaptively transfer relational information between class representations. DART is the first framework, to our knowledge, to explicitly integrate external LLM-derived relational knowledge for adaptive inter-class transfer while simultaneously performing adaptive intra-class refinement under weak supervision for OV-MLR. Extensive experiments on challenging benchmarks demonstrate that our DART achieves new state-of-the-art performance, validating its effectiveness.

3D affordance reasoning plays a critical role in associating human instructions with the functional regions of 3D objects, facilitating precise, task-oriented manipulations in embodied AI. However, current methods, which predominantly depend on sparse 3D point clouds, exhibit limited generalizability and robustness due to their sensitivity to coordinate variations and the inherent sparsity of the data. By contrast, 3D Gaussian Splatting (3DGS) delivers high-fidelity, real-time rendering with minimal computational overhead by representing scenes as dense, continuous distributions. This positions 3DGS as a highly effective approach for capturing fine-grained affordance details and improving recognition accuracy. Nevertheless, its full potential remains largely untapped due to the absence of large-scale, 3DGS-specific affordance datasets. To overcome these limitations, we present 3DAffordSplat, the first large-scale, multi-modal dataset tailored for 3DGS-based affordance reasoning. This dataset includes 23,672 Gaussian instances, 8,231 point cloud instances, and 6,631 manually annotated affordance labels, encompassing 21 object categories and 18 affordance types. Building upon this dataset, we introduce AffordSplatNet, a novel model specifically designed for affordance reasoning using 3DGS representations. AffordSplatNet features an innovative cross-modal structure alignment module that exploits structural consistency priors to align 3D point cloud and 3DGS representations, resulting in enhanced affordance recognition accuracy. Extensive experiments demonstrate that the 3DAffordSplat dataset significantly advances affordance learning within the 3DGS domain, while AffordSplatNet consistently outperforms existing methods across both seen and unseen settings, highlighting its robust generalization capabilities. Code, model, and video are available at https://hcplab-sysu.github.io/3DAffordSplat.

Instruction-Tuning (IT) was recently found the impressive data efficiency in post-training large language models (LLMs). While the pursuit of efficiency predominantly focuses on sequence-level curation, often overlooking the nuanced impact of critical tokens and the inherent risks of token noise and biases. Drawing inspiration from bi-level coreset selection, our work provides the principled view of the motivation behind selecting instructions' responses. It leads to our approach Quadratic Coreset Selection (QCS) that reconciles sequence-level and token-level influence contributions, deriving more expressive LLMs with established theoretical result. Despite the original QCS framework challenged by prohibitive computation from inverted LLM-scale Hessian matrices, we overcome this barrier by proposing a novel QCS probabilistic variant, which relaxes the original formulation through re-parameterized densities. This innovative solver is efficiently learned using hierarchical policy gradients without requiring back-propagation, achieving provable convergence and certified asymptotic equivalence to the original objective. Our experiments demonstrate QCS's superior sequence-level data efficiency and reveal how strategically leveraging token-level influence elevates the performance ceiling of data-efficient IT. Furthermore, QCS's adaptability is showcased through its successes in regular IT and challenging targeted IT scenarios, particularly in the cases of free-form complex instruction-following and CoT reasoning. They underscore QCS's potential for a wide array of versatile post-training applications.

Egocentric Referring Video Object Segmentation (Ego-RVOS) aims to segment the specific object actively involved in a human action, as described by a language query, within first-person videos. This task is critical for understanding egocentric human behavior. However, achieving such segmentation robustly is challenging due to ambiguities inherent in egocentric videos and biases present in training data. Consequently, existing methods often struggle, learning spurious correlations from skewed object-action pairings in datasets and fundamental visual confounding factors of the egocentric perspective, such as rapid motion and frequent occlusions. To address these limitations, we introduce Causal Ego-REferring Segmentation (CERES), a plug-in causal framework that adapts strong, pre-trained RVOS backbones to the egocentric domain. CERES implements dual-modal causal intervention: applying backdoor adjustment principles to counteract language representation biases learned from dataset statistics, and leveraging front-door adjustment concepts to address visual confounding by intelligently integrating semantic visual features with geometric depth information guided by causal principles, creating representations more robust to egocentric distortions. Extensive experiments demonstrate that CERES achieves state-of-the-art performance on Ego-RVOS benchmarks, highlighting the potential of applying causal reasoning to build more reliable models for broader egocentric video understanding.

To improve detection robustness in adverse conditions (e.g., haze and low light), image restoration is commonly applied as a pre-processing step to enhance image quality for the detector. However, the functional mismatch between restoration and detection networks can introduce instability and hinder effective integration---an issue that remains underexplored. We revisit this limitation through the lens of Lipschitz continuity, analyzing the functional differences between restoration and detection networks in both the input space and the parameter space. Our analysis shows that restoration networks perform smooth, continuous transformations, while object detectors operate with discontinuous decision boundaries, making them highly sensitive to minor perturbations. This mismatch introduces instability in traditional cascade frameworks, where even imperceptible noise from restoration is amplified during detection, disrupting gradient flow and hindering optimization. To address this, we propose Lipschitz-regularized object detection (LROD), a simple yet effective framework that integrates image restoration directly into the detector’s feature learning, harmonizing the Lipschitz continuity of both tasks during training. We implement this framework as Lipschitz-regularized YOLO (LR-YOLO), extending seamlessly to existing YOLO detectors. Extensive experiments on haze and low-light benchmarks demonstrate that LR-YOLO consistently improves detection stability, optimization smoothness, and overall accuracy.

Recently, numerous benchmarks have been developed to evaluate the logical reasoning abilities of large language models (LLMs). However, assessing the equally important creative capabilities of LLMs is challenging due to the subjective, diverse, and data-scarce nature of creativity, especially in multimodal scenarios. In this paper, we consider the comprehensive pipeline for evaluating the creativity of multimodal LLMs, with a focus on suitable evaluation platforms and methodologies. First, we find the Oogiri game-a creativity-driven task requiring humor, associative thinking, and the ability to produce unexpected responses to text, images, or both. This game aligns well with the input-output structure of modern multimodal LLMs and benefits from a rich repository of high-quality, human-annotated creative responses, making it an ideal platform for studying LLM creativity. Next, beyond using the Oogiri game for standard evaluations like ranking and selection, we propose LoTbench, an interactive, causality-aware evaluation framework, to further address some intrinsic risks in standard evaluations, such as information leakage and limited interpretability. The proposed LoTbench not only quantifies LLM creativity more effectively but also visualizes the underlying creative thought processes. Our results show that while most LLMs exhibit constrained creativity, the performance gap between LLMs and humans is not insurmountable. Furthermore, we observe a strong correlation between results from the multimodal cognition benchmark MMMU and LoTbench, but only a weak connection with traditional creativity metrics. This suggests that LoTbench better aligns with human cognitive theories, highlighting cognition as a critical foundation in the early stages of creativity and enabling the bridging of diverse concepts. Project Page.

Infrared and visible image fusion (IVIF) is increasingly applied in critical fields such as video surveillance and autonomous driving systems. Significant progress has been made in deep learning-based fusion methods. However, these models frequently encounter out-of-distribution (OOD) scenes in real-world applications, which severely impact their performance and reliability. Therefore, addressing the challenge of OOD data is crucial for the safe deployment of these models in open-world environments. Unlike existing research, our focus is on the challenges posed by OOD data in real-world applications and on enhancing the robustness and generalization of models. In this paper, we propose an infrared-visible fusion framework based on Multi-View Augmentation. For external data augmentation, Top-k Selective Vision Alignment is employed to mitigate distribution shifts between datasets by performing RGB-wise transformations on visible images. This strategy effectively introduces augmented samples, enhancing the adaptability of the model to complex real-world scenarios. Additionally, for internal data augmentation, self-supervised learning is established using Weak-Aggressive Augmentation. This enables the model to learn more robust and general feature representations during the fusion process, thereby improving robustness and generalization. Extensive experiments demonstrate that the proposed method exhibits superior performance and robustness across various conditions and environments. Our approach significantly enhances the reliability and stability of IVIF tasks in practical applications.

Infrared small target detection (ISTD) is highly sensitive to sensor type, observation conditions, and the intrinsic properties of the target. These factors can introduce substantial variations in the distribution of acquired infrared image data, a phenomenon known as domain shift. Such distribution discrepancies significantly hinder the generalization capability of ISTD models across diverse scenarios. To tackle this challenge, this paper introduces an ISTD framework enhanced by domain adaptation. To alleviate distribution shift between datasets and achieve cross-sample alignment, we introduce Cross-view Channel Alignment (CCA). Additionally, we propose the Cross-view Top-K Fusion strategy, which integrates target information with diverse background features, enhancing the model’s ability to extract critical data characteristics. To further mitigate the impact of noise on ISTD, we develop a Noise-guided Representation learning strategy. This approach enables the model to learn more noise-resistant feature representations, to improve its generalization capability across diverse noisy domains. Finally, we develope a dedicated infrared small target dataset, RealScene-ISTD. Compared to state-of-the-art methods, our approach demonstrates superior performance in terms of detection probability (Pd), false alarm rate (Fa), and intersection over union (IoU). The code is available at: https://github.com/luy0222/RealScene-ISTD.

Radiology Report Generation (RRG) is essential for computer-aided diagnosis and medication guidance, which can relieve the heavy burden of radiologists by automatically generating the corresponding radiology reports according to the given radiology image. However, generating accurate lesion descriptions remains challenging due to spurious correlations from visual-linguistic biases and inherent limitations of radiological imaging, such as low resolution and noise interference. To address these issues, we propose a two-stage framework named Cross-Modal Causal Representation Learning (CMCRL), consisting of the Radiological Cross-modal Alignment and Reconstruction Enhanced (RadCARE) pre-training and the Visual-Linguistic Causal Intervention (VLCI) fine-tuning. In the pre-training stage, RadCARE introduces a degradation-aware masked image restoration strategy tailored for radiological images, which reconstructs high-resolution patches from low-resolution inputs to mitigate noise and detail loss. Combined with a multiway architecture and four adaptive training strategies (e.g., text postfix generation with degraded images and text prefixes), RadCARE establishes robust cross-modal correlations even with incomplete data. In the VLCI phase, we deploy causal front-door intervention through two modules: the Visual Deconfounding Module (VDM) disentangles local-global features without fine-grained annotations, while the Linguistic Deconfounding Module (LDM) eliminates context bias without external terminology databases. Experiments on IU-Xray and MIMIC-CXR show that our CMCRL pipeline significantly outperforms state-of-the-art methods, with ablation studies confirming the necessity of both stages. Code and models are available at https://github.com/WissingChen/CMCRL.

Maintaining stylistic consistency is crucial for the cohesion and aesthetic appeal of images, a fundamental requirement in effective image editing and inpainting. However, existing methods primarily focus on the semantic control of generated content, often neglecting the critical task of preserving this consistency. In this work, we introduce the Neural Scene Designer (NSD), a novel framework that enables photo-realistic manipulation of user-specified scene regions while ensuring both semantic alignment with user intent and stylistic consistency with the surrounding environment. NSD leverages an advanced diffusion model, incorporating two parallel cross-attention mechanisms that separately process text and style information to achieve the dual objectives of semantic control and style consistency. To capture fine-grained style representations, we propose the Progressive Self-style Representational Learning (PSRL) module. This module is predicated on the intuitive premise that different regions within a single image share a consistent style, whereas regions from different images exhibit distinct styles. The PSRL module employs a style contrastive loss that encourages high similarity between representations from the same image while enforcing dissimilarity between those from different images. Furthermore, to address the lack of standardized evaluation protocols for this task, we establish a comprehensive benchmark. This benchmark includes competing algorithms, dedicated style-related metrics, and diverse datasets and settings to facilitate fair comparisons. Extensive experiments conducted on our benchmark demonstrate the effectiveness of the proposed framework.

The class-agnostic counting (CAC) task has recently been proposed to solve the problem of counting all objects of an arbitrary class with several exemplars given in the input image. To address this challenging task, existing leading methods all resort to density map regression, which renders them impractical for downstream tasks that require object locations and restricts their ability to well explore the scale information of exemplars for supervision. Meanwhile, they generally model the interaction between the input image and the exemplars in an exemplar-by-exemplar way, which is inefficient and may not fully synthesize information from all exemplars. To address these limitations, we propose a novel localization-based CAC approach, termed Scale-modulated Query and Localization Network (SQLNet). It fully explores the scales of exemplars in both the query and localization stages and achieves effective counting by accurately locating each object and predicting its approximate size. Specifically, during the query stage, rich discriminative representations of the target class are acquired by the Hierarchical Exemplars Collaborative Enhancement (HECE) module from the few exemplars through multi-scale exemplar cooperation with equifrequent size prompt embedding. These representations are then fed into the Exemplars-Unified Query Correlation (EUQC) module to interact with the query features in a unified manner and produce the correlated query tensor. In the localization stage, the Scale-aware Multi-head Localization (SAML) module utilizes the query tensor to predict the confidence, location, and size of each potential object. Moreover, a scale-aware localization loss is introduced, which exploits flexible location associations and exemplar scales for supervision to optimize the model performance. Extensive experiments demonstrate that SQLNet outperforms state-of-the-art methods on popular CAC benchmarks, achieving excellent performance not only in counting accuracy but also in localization and bounding box generation.

Metro Origin-Destination (OD) prediction is a crucial yet challenging spatial-temporal prediction task in urban computing, which aims to accurately forecast cross-station ridership for optimizing metro scheduling and enhancing overall transport efficiency. Analyzing fine-grained and comprehensive relations among stations effectively is imperative for metro OD prediction. However, existing metro OD models either mix information from multiple OD pairs from the station's perspective or exclusively focus on a subset of OD pairs. These approaches may overlook fine-grained relations among OD pairs, leading to difficulties in predicting potential anomalous conditions. To address these challenges, we learn traffic evolution from the perspective of all OD pairs and propose a fine-grained spatialtemporal MLP architecture for metro OD prediction, namely ODMixer. Specifically, our ODMixer has double-branch structure and involves the Channel Mixer, the Multi-view Mixer, and the Bidirectional Trend Learner. The Channel Mixer aims to capture short-term temporal relations among OD pairs, the Multi-view Mixer concentrates on capturing spatial relations from both origin and destination perspectives. To model long-term temporal relations, we introduce the Bidirectional Trend Learner. Extensive experiments on two large-scale metro OD prediction datasets HZMOD and SHMO demonstrate the advantages of our ODMixer. Our code is available at https://github.com/KLatitude/ODMixer

Deep reinforcement learning (RL) has witnessed remarkable success in a wide range of control tasks. To overcome RL's notorious sample inefficiency, prior studies have explored data augmentation techniques leveraging collected transition data. However, these methods face challenges in synthesizing transitions adhering to the authentic environment dynamics, especially when the transition is high-dimensional and includes many redundant/irrelevant features to the task. In this article, we introduce continuous value assignment (CVA), an innovative optimization-level data augmentation approach that directly synthesizes novel training data in the state-action value space, effectively bypassing the need for explicit transition modeling. The key intuition of our method is that the transition plays an intermediate role in calculating the state-action value during optimization, and therefore directly augmenting the state-action value is more causally related to the optimization process. Specifically, our CVA combines parameterized value prediction and nonparametric value interpolation from neighboring states, resulting in doubly robust target values w.r.t. novel states and actions. Extensive experiments demonstrate CVA's substantial improvements in sample efficiency across complex continuous control tasks, surpassing several advanced baselines.

Accurate inference of fine-grained traffic flow from coarse-grained one is an emerging yet crucial problem, which can help greatly reduce the number of the required traffic monitoring sensors for cost savings. In this work, we note that traffic flow has a high correlation with road network, which was either completely ignored or simply treated as an external factor in previous works. To facilitate this problem, we propose a novel road-aware traffic flow magnifier (RATFM) that explicitly exploits the prior knowledge of road networks to fully learn the road-aware spatial distribution of fine-grained traffic flow. Specifically, a multidirectional 1-D convolutional layer is first introduced to extract the semantic feature of the road network. Subsequently, we incorporate the road network feature and coarse-grained flow feature to regularize the short-range spatial distribution modeling of road-relative traffic flow. Furthermore, we take the road network feature as a query to capture the long-range spatial distribution of traffic flow with a transformer architecture. Benefiting from the road-aware inference mechanism, our method can generate high-quality fine-grained traffic flow maps. Extensive experiments on three real-world datasets show that the proposed RATFM outperforms state-of-the-art models under various scenarios. Our code and datasets are released at https://github.com/luimoli/RATFM.

We present a novel data-driven Parametric Linear Blend Skinning (PLBS) model meticulously crafted for generalized 3D garment dressing and animation. Previous data-driven methods are impeded by certain challenges including overreliance on human body modeling and limited adaptability across different garment shapes. Our method resolves these challenges via two goals: 1) Develop a model based on garment modeling rather than human body modeling. 2) Separately construct low-dimensional sub-spaces for modeling in-plane deformation (such as variation in garment shape and size) and out-of-plane deformation (such as deformation due to varied body size and motion). Therefore, we formulate garment deformation as a PLBS model controlled by canonical 3D garment mesh, vertex-based skinning weights and associated local patch transformation. Unlike traditional LBS models specialized for individual objects, PLBS model is capable of uniformly expressing varied garments and bodies, the in-plane deformation is encoded on the canonical 3D garment and the out-of-plane deformation is controlled by the local patch transformation. Besides, we propose novel 3D garment registration and skinning weight decomposition strategies to obtain adequate data to build PLBS model under different garment categories. Furthermore, we employ dynamic fine-tuning to complement high-frequency signals missing from LBS for unseen testing data. Experiments illustrate that our method is capable of modeling dynamics for loose-fitting garments, outperforming previous data-driven modeling methods using different sub-space modeling strategies. We showcase that our method can factorize and be generalized for varied body sizes, garment shapes, garment sizes and human motions under different garment categories.

Building scalable vision-language models to learn from diverse, multimodal data remains an open challenge. In this paper, we introduce an Efficient Vision-languagE foundation model, namely EVE, which is one unified multimodal Transformer pre-trained solely by one unified pre-training task. Specifically, EVE encodes both vision and language within a shared Transformer network integrated with modality-aware sparse Mixture-of-Experts (MoE) modules, which capture modality-specific information by selectively switching to different experts. To unify pre-training tasks of vision and language, EVE performs masked signal modeling on image-text pairs to reconstruct masked signals, i.e., image pixels and text tokens, given visible signals. This simple yet effective pre-training objective accelerates training by 4x compared to the model pre-trained with Image-Text Contrastive and Image-Text Matching losses. Owing to the combination of the unified architecture and pre-training task, EVE is easy to scale up, enabling better downstream performance with fewer resources and faster training speed. Despite its simplicity, EVE achieves state-of-the-art performance on various vision-language downstream tasks, including visual question answering, visual reasoning, and image-text retrieval.

Invariant representation learning (IRL) encourages the prediction from invariant causal features to labels deconfounded from the environments, advancing the technical roadmap of out-of-distribution (OOD) generalization. Despite spotlights around, recent theoretical result verified that some causal features recovered by IRLs merely pretend domain-invariantly in the training environments but fail in unseen domains. The fake invariance severely endangers OOD generalization since the trustful objective can not be diagnosed and existing causal remedies are invalid to rectify. In this paper, we review a IRL family (InvRat) under the Partially and Fully Informative Invariant Feature Structural Causal Models (PIIF SCM /FIIF SCM) respectively, to certify their weaknesses in representing fake invariant features, then, unify their causal diagrams to propose ReStructured SCM (RS-SCM). RS-SCM can ideally rebuild the spurious and the fake invariant features simultaneously. Given this, we further develop an approach based on conditional mutual information with respect to RS-SCM, then rigorously rectify the spurious and fake invariant effects. It can be easily implemented by a small feature selection subnet introduced in the IRL family, which is alternatively optimized to achieve our goal. Experiments verified the superiority of our approach to fight against the fake invariant issue across a variety of OOD generalization benchmarks.

The filter bubble is a notorious issue in Recommender Systems (RSs), which describes the phenomenon whereby users are exposed to a limited and narrow range of information or content that reinforces their existing dominant preferences and beliefs. This results in a lack of exposure to diverse and varied content. Many existing works have predominantly examined filter bubbles in static or relatively-static recommendation settings. However, filter bubbles will be continuously intensified over time due to the feedback loop between the user and the system in the real-world online recommendation. To address these issues, we propose a novel paradigm, Multi-Facet Preference Learning for Pricking Filter Bubbles in Conversational Recommender System (FacetCRS), which aims to burst filter bubbles in the conversational recommender system (CRS) through timely user-item interactions via natural language conversations. By considering diverse user preferences and intentions, FacetCRS automatically model user preference into multi-facets, including entity-, word-, context-, and review-facet, to capture diverse and dynamic user preferences to prick filter bubbles in the CRS. It is an end-to-end CRS framework to adaptively learn representations of various levels of preference facet and diverse types of external knowledge. Extensive experiments on two publicly available benchmark datasets demonstrate that our proposed method achieves state-of-the-art performance in mitigating filter bubbles and enhancing recommendation quality in CRS.

Yongsen Zheng, Ruilin Xu, Ziliang Chen, Guohua Wang, Mingjie Qian, Jinghui Qin, Liang Lin. Proceedings of the 62nd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers). 2024.

Speech-preserving facial expression manipulation (SPFEM) aims to modify facial emotions while meticulously maintaining the mouth animation associated with spoken content. Current works depend on inaccessible paired training samples for the person, where two aligned frames exhibit the same speech content yet differ in emotional expression, limiting the SPFEM applications in real-world scenarios. In this work, we discover that speak-ers who convey the same content with different emotions exhibit highly correlated local facial animations, providing valuable supervision for SPFEM. To capitalize on this insight, we propose a novel adaptive spatial coherent correlation learning (ASCCL) algorithm, which models the aforementioned correlation as an explicit metric and integrates the metric to supervise manipulating facial expression and meanwhile better preserving the facial animation of spoken contents. To this end, it first learns a spatial coherent correlation metric, ensuring the visual disparities of adjacent local regions of the image belonging to one emotion are similar to those of the corresponding counterpart of the image belonging to another emotion. Recognizing that visual disparities are not uniform across all regions, we have also crafted a disparity-aware adaptive strategy that prioritizes regions that present greater challenges. During SPFEM model training, we construct the adaptive spatial coherent correlation metric between corresponding local regions of the input and output images as addition loss to supervise the generation process. We conduct extensive experiments on variant datasets, and the results demonstrate the effectiveness of the proposed ASCCL algorithm. Code is publicly available at https://githiub.com/jianmanlincjx/ASCCL

Open vocabulary object detection (OVD) aims at seeking an optimal object detector capable of recognizing objects from both base and novel categories. Recent advances leverage knowledge distillation to transfer insightful knowledge from pre-trained large-scale vision-language models to the task of object detection, significantly generalizing the powerful capabilities of the detector to identify more unknown object categories. However, these methods face significant challenges in background interpretation and model overfitting and thus often result in the loss of crucial back-ground knowledge, giving rise to sub-optimal inference performance of the detector. To mitigate these issues, we present a novel OVD framework termed LBP to propose learning background prompts to harness explored implicit background knowledge, thus enhancing the detection performance w.r.t. base and novel categories. Specifically, we devise three modules: Background Category-specific Prompt, Background Object Discovery, and Inference Probability Rectification, to empower the detector to discover, represent, and leverage implicit object knowledge explored from background proposals. Evaluation on two benchmark datasets, OV-COCO and OV-LVIS, demonstrates the superiority of our proposed method over existing state-of-the-art approaches in handling the OVD tasks.

Neural implicit fields have been a de facto standard in novel view synthesis. Recently, there exist some methods exploring fusing multiple modalities within a single field, aiming to share implicit features from different modalities to enhance reconstruction performance. However, these modalities often exhibit misaligned behaviors: optimizing for one modality, such as LiDAR, can adversely affect another, like camera performance, and vice versa. In this work, we conduct comprehensive analyses on the multimodal implicit field of LiDAR-camera joint synthesis, revealing the underlying issue lies in the misalignment of different sensors. Furthermore, we introduce AlignMiF, a geometrically aligned multimodal implicit field with two proposed modules: Geometry-Aware Alignment (GAA) and Shared Geometry Initialization (SGI). These modules effectively align the coarse geometry across different modalities, significantly enhancing the fusion process between LiDAR and camera data. Through extensive experiments across various datasets and scenes, we demonstrate the effectiveness of our approach in facilitating better interaction between LiDAR and camera modalities within a unified neural field. Specifically, our proposed AlignMiF, achieves remarkable improvement over recent implicit fusion methods (+2.01 and +3.11 image PSNR on the KITTI-360 and Waymo datasets) and consistently surpasses single modality performance (13.8% and 14.2% reduction in LiDAR Cham-fer Distance on the respective datasets). Code release: https://github.com/tangtaogo/alignmif.

Chain-of-Thought (CoT) [2], [3] guides large language models (LLMs) to reason step-by-step, and can motivate their logical reasoning ability. While effective for logi-cal tasks, CoT is not conducive to creative problem-solving which often requires out-of-box thoughts and is crucial for innovation advancements. In this paper, we explore the Leap-of-Thought (LoT) abilities within LLMs - a non-sequential, creative paradigm involving strong associations and knowledge leaps. To this end, we study LLMs on the popular Oogiri game which needs participants to have good creativity and strong associative thinking for responding unexpectedly and humorously to the given image, text, or both, and thus is suitable for LoT study. Then to investi-gate LLMs' LoT ability in the Oogiri game, we first build a multimodal and multilingual Oogiri-GO dataset which contains over 130,000 samples from the Oogiri game, and observe the insufficient LoT ability or failures of most existing LLMs on the Oogiri game. Accordingly, we introduce a creative Leap-of-Thought (CLoT) paradigm to improve LLM's LoT ability. CLoT first formulates the Oogiri-GO dataset into LoT-oriented instruction tuning data to train pre-trained LLM for achieving certain LoT humor generation and discrimination abilities. Then CLoT designs an ex-plorative self-refinement that encourages the LLM to gener-ate more creative LoT data via exploring parallels between seemingly unrelated concepts and selects high-quality data to train itself for self-refinement. CLoT not only excels in humor generation in the Oogiri game as shown in Fig. 1 but also boosts creative abilities in various tasks like “cloud guessing game” and “divergent association task”. These findings advance our understanding and offer a pathway to improve LLMs' creative capacities for innovative applications across domains. The dataset, code, and models have been released online: https://zhongshsh.github.io/CLoT.

The Matthew effect is a big challenge in Recommender Systems (RSs), where popular items tend to receive increasing attention, while less popular ones are often overlooked, perpetuating existing disparities.Although many existing methods attempt to mitigate Matthew effect in the static or quasi-static recommendation scenarios, such issue will be more pronounced as users engage with the system over time.To this end, we propose a novel framework, Multi-Hypergraph Boosted Multi-Interest Self-Supervised Learning for Conversational Recommendation (HiCore), aiming to address Matthew effect in the Conversational Recommender System (CRS) involving the dynamic user-system feedback loop.It devotes to learn multi-level user interests by building a set of hypergraphs (i.e., item-, entity-, word-oriented multiple-channel hypergraphs) to alleviate the Matthew effec.Extensive experiments on four CRS-based datasets showcase that HiCore attains a new state-of-the-art performance, underscoring its superiority in mitigating the Matthew effect effectively.Our code is available at https://github.com/zysensmile/HiCore.

Semi-Supervised Object Detection (SSOD) has achieved resounding success by leveraging unlabeled data to improve detection performance. However, in Open Scene Semi-Supervised Object Detection (O-SSOD), unlabeled data may contains unknown objects not observed in the labeled data, which will increase uncertainty in the model’s predictions for known objects. It is detrimental to the current methods that mainly rely on self-training, as more uncertainty leads to the lower localization and classification precision of pseudo labels. To this end, we propose Credible Teacher, an end-to-end framework. Credible Teacher adopts an interactive teaching mechanism using flexible labels to prevent uncertain pseudo labels from misleading the model and gradually reduces its uncertainty through the guidance of other credible pseudo labels. Empirical results have demonstrated our method effectively restrains the adverse effect caused by O-SSOD and significantly outperforms existing counterparts.

We propose the attention-inspired numerical solver (AttNS), a concise method that helps the generalization and robustness issues faced by the AI-Hybrid numerical solver in solving differential equations due to limited data. AttNS is inspired by the effectiveness of attention modules in Residual Neural Networks (ResNet) in enhancing model generalization and robustness for conventional deep learning tasks. Drawing from the dynamical system perspective of ResNet, We seamlessly incorporate attention mechanisms into the design of numerical methods tailored for the characteristics of solving differential equations. Our results on benchmarks, ranging from high-dimensional problems to chaotic systems, showcase AttNS consistently enhancing various numerical solvers without any intricate model crafting. Finally, we analyze AttNS experimentally and theoretically, demonstrating its ability to achieve strong generalization and robustness while ensuring the convergence of the solver. This includes requiring less data compared to other advanced methods to achieve comparable generalization errors and better prevention of numerical explosion issues when solving differential equations.

A codebook designed for learning discrete distributions in latent space has demonstrated state-of-the-art results on generation tasks. This inspires us to explore what distribution of codebook is better. Following the spirit of Kepler's Conjecture, we cast the codebook training as solving the sphere packing problem and derive a Kepler codebook with a compact and structured distribution to obtain a codebook for image representations. Furthermore, we implement the Kepler codebook training by simply employing this derived distribution as regularization and using the codebook partition method. We conduct extensive experiments to evaluate our trained codebook for image reconstruction and generation on natural and human face datasets, respectively, achieving significant performance improvement. Besides, our Kepler codebook has demonstrated superior performance when evaluated across datasets and even for reconstructing images with different resolutions. Our trained models and source codes will be publicly released.

Speech-preserving Facial Expression Manipulation (SPFEM) aims to alter facial emotions in video content while preserving the facial movements associated with speech. Current works often fall short due to the inadequate representation of emotion as well as the absence of time-aligned paired data-two corresponding frames from the same speaker that showcase the same speech content but differ in emotional expression. In this work, we introduce a novel framework, Self-Supervised Emotion Representation Disentanglement (SSERD), to disentangle emotion representation for accurate emotion transfer while implementing a paired data construction module to facilitate automated, photorealistic facial animations. Specifically, We developed a module for learning emotion latent codes using StyleGAN's latent space, employing a cross-attention mechanism to extract and predict emotion editing codes, with contrastive learning to differentiate emotions. To overcome the lack of strictly paired data in the SPFEM task, we exploit pretrained StyleGAN to generate paired data, focusing on expression vectors unrelated to mouth shape. Additionally, we employed a hybrid training strategy using both synthetic paired and real unpaired data to enhance the realism of SPFEM model's generated images. Extensive experiments conducted on benchmark datasets, including MEAD and RAVDESS, have validated the effectiveness of our framework, demonstrating its superior capability in generating photorealistic and expressive facial animations.

Diversity plays a crucial role in Recommender Systems (RSs) as it ensures a wide range of recommended items, providing users with access to new and varied options. Without diversity, users often encounter repetitive content, limiting their exposure to novel choices. While significant efforts begin to enhance recommendation diversification in static offline scenarios, relatively less attention has been given to online Conversational Recommender Systems (CRSs). However, the lack of recommendation diversity in CRSs will increasingly exacerbate over time due to the dynamic user-system feedback loop, resulting in challenges such as the Matthew effect, filter bubbles, and echo chambers. To address these issues, we propose a novel paradigm, User-Centric Multi-Interest Learning for Conversational Movie Recommendation (CoMoRec), aiming to learn multiple user interests to improve result diversity for movie recommendations. Firstly, CoMoRec automatically models various facets of user interests, including context-, graph-, and review-based interests, to explore a wide range of user potential intentions. Then, it leverages these multi-aspect user interests to accurately predict personalized and diverse movie recommendations and generate fluent and informative responses during conversations. Extensive experiments on two publicly CRS-based movie datasets show that our CoMoRec achieves a new state-of-the-art performance and the superiority of improving recommendation diversity in the CRS.

Multimodal recommender systems utilize various types of information to model user preferences and item features, helping users discover items aligned with their interests. The integration of multimodal information mitigates the inherent challenges in recommender systems, e.g., the data sparsity problem and cold-start issues. However, it simultaneously magnifies certain risks from multimodal information inputs, such as information adjustment risk and inherent noise risk. These risks pose crucial challenges to the robustness of recommendation models. In this paper, we analyze multimodal recommender systems from the novel perspective of flat local minima and propose a concise yet effective gradient strategy called Mirror Gradient (MG). This strategy can implicitly enhance the model's robustness during the optimization process, mitigating instability risks arising from multimodal information inputs. We also provide strong theoretical evidence and conduct extensive empirical experiments to show the superiority of MG across various multimodal recommendation models and benchmarks. Furthermore, we find that the proposed MG can complement existing robust training methods and be easily extended to diverse advanced recommendation models, making it a promising new and fundamental paradigm for training multimodal recommender systems. The code is released at https://github.com/Qrange-group/Mirror-Gradient.

Neural Architecture Search (NAS), aiming at automatically designing neural architectures by machines, has been considered a key step toward automatic machine learning. One notable NAS branch is the weight-sharing NAS, which significantly improves search efficiency and allows NAS algorithms to run on ordinary computers. Despite receiving high expectations, this category of methods suffers from low search effectiveness. By employing a generalization boundedness tool, we demonstrate that the devil behind this drawback is the untrustworthy architecture rating with the oversized search space of the possible architectures. Addressing this problem, we modularize a large search space into blocks with small search spaces and develop a family of models with the distilling neural architecture (DNA) techniques. These proposed models, namely a DNA family, are capable of resolving multiple dilemmas of the weight-sharing NAS, such as scalability, efficiency, and multi-modal compatibility. Our proposed DNA models can rate all architecture candidates, as opposed to previous works that can only access a sub- search space using heuristic algorithms. Moreover, under a certain computational complexity constraint, our method can seek architectures with different depths and widths. Extensive experimental evaluations show that our models achieve state-of-the-art top-1 accuracy of 78.9% and 83.6% on ImageNet for a mobile convolutional network and a small vision transformer, respectively. Additionally, we provide in-depth empirical analysis and insights into neural architecture ratings.

Recently, text-to-image diffusion models become a new paradigm in image processing fields, including content generation, image restoration and image-to-image translation. Given a target prompt, Denoising Diffusion Probabilistic Models (DDPM) are able to generate realistic yet eligible images. With this appealing property, the image translation task has the potential to be free from target image samples for supervision. By using a target text prompt for domain adaption, the diffusion model is able to implement zero-shot image-to-image translation advantageously. However, the sampling and inversion processes of DDPM are stochastic, and thus the inversion process often fail to reconstruct the input content. Specifically, the displacement effect will gradually accumulated during the diffusion and inversion processes, which led to the reconstructed results deviating from the source domain. To make reconstruction explicit, we propose a prompt redescription strategy to realize a mirror effect between the source and reconstructed image in the diffusion model (MirrorDiffusion). More specifically, a prompt redescription mechanism is investigated to align the text prompts with latent code at each time step of the Denoising Diffusion Implicit Models (DDIM) inversion to pursue a structure-preserving reconstruction. With the revised DDIM inversion, MirrorDiffusion is able to realize accurate zero-shot image translation by editing optimized text prompts and latent code. Extensive experiments demonstrate that MirrorDiffusion achieves superior performance over the state-of-the-art methods on zero-shot image translation benchmarks by clear margins and practical model stability.

Though quite challenging, training a deep neural network for automatically solving Math Word Problems (MWPs) has increasingly attracted attention due to its significance in investigating how a machine can understand and reason complex problems like a human. However, the data volume of existing high-quality MWP datasets is far from sufficient to train a robust solver since collecting these datasets would cost a very high price, i.e., they require professional knowledge that accords with the educational standard and massive accessible data. This data bottleneck inspires us to consider using cost-effective data augmentation methods to improve the utilization of the existing data and enhance the performance of an MWP solver. Nevertheless, the traditional input-based data augmentation methods for training natural image or language models are incompetent for training MWP solvers due to the following two reasons. First, MWPs are concise yet comprehensive, so these data augmentation methods are prone to make them more ambiguous. Second, the mathematical dependencies grounded in the problems must be maintained when a batch of augmented examples is generated during the data augmentation process. To address these issues, we propose a simple yet effective data augmentation method called the Introspective Data Augmentation Method (IDAM) that allows the MWP examples to be augmented in latent space during the training of the neural network, instead of making perturbations over the input data. In particular, our IDAM is capable of applying different data augmentation operations on the latent feature representations of MWPs to produce new examples. Moreover, a new training objective is developed to constrain the mathematical dependency consistency between the original MWP and the produced ones. Extensive experiments conducted on standard benchmarks demonstrate the effectiveness of IDAM in generally improving the performance of existing MWP solvers without any elaborated model crafting.

Dynamic hand gesture is an emerging and promising biometric trait containing both physiological and behavioral characteristics. Possessing the two kinds of characteristics makes dynamic hand gesture have more identity information enabling more accurate and secure authentication theoretically, but also poses a challenge of efficient fine-grained spatiotemporal feature extraction. This challenge involves a seemingly paradoxical problem that high-frame-rate videos are required for behavioral characteristic analysis, but they can also introduce high computational costs. To mitigate this issue, we propose a Frequency Spatiotemporal Attention Network (FSTA-Net) with a focus on satisfying the high-performance and low-computation requirements of authentication systems. The FSTA-Net is established with a two-stage identity characteristic analysis paradigm for short- and long-term modeling. Specifically, considering that models prefer to analyze physiological characteristics which are relatively straightforward to understand, we first design a Behavior Enhanced (BE) module to emphasize hand motions and reduce redundant information to facilitate local identity feature distillation in the first stage. We then present a Frequency Spatiotemporal Attention (FSTA) module to summarize global identity features with decent FLOPs and GPU memory occupation in the second stage. Incorporating the BE and FSTA modules enables them to complement each other’s strengths, resulting in a clear-cut improvement in equal error rate and running speed. Extensive experiments on the SCUT-DHGA dataset demonstrate the superiority of the FSTA-Net. The code is available athttps://github.com/SCUT-BIP-Lab/FSTA-Net.

Given a descriptive text query, text-based person search (TBPS) aims to retrieve the best matched target person from an image gallery. Such a cross-modal retrieval task is quite challenging due to significant modality gap, fine-grained differences and insufficiency of annotated data. To better align the two modalities, most existing works focus on introducing sophisticated network structures and auxiliary tasks, which are complex and hard to implement. In this paper, we propose a simple yet effective dual Transformer model for text-based person search. By exploiting a hardness-aware contrastive learning strategy, our model achieves state-of-the-art performance without any special design for local feature alignment or side information. Moreover, we propose a proximity data generation (PDG) module to automatically produce more diverse data for cross-modal training. The PDG module first introduces an automatic generation algorithm based on a text-to-image diffusion model, which generates new text-image pair samples in the proximity space of original ones. Then it combines approximate text generation and feature-level mixup during training to further strengthen the data diversity. The PDG module can largely guarantee the reasonability of the generated samples that are directly used for training without any human inspection for noise rejection. It improves the performance of our model significantly, providing a feasible solution to the data insufficiency problem faced by such fine-grained visual-linguistic tasks. Extensive experiments on two popular datasets of the TBPS task (i.e., CUHK-PEDES and ICFG-PEDES) show that the proposed approach outperforms state-of-the-art approaches evidently, e.g., improving by 3.88%, 4.02%, 2.92% in terms of Top1, Top5, Top10 on CUHK-PEDES.

Single-frame infrared small target (SIRST) detection aims to recognize small targets from clutter backgrounds. Recently, convolutional neural networks have achieved significant advantages in general object detection. With the development of Transformer, the scale of SIRST models is constantly increasing. Due to the limited training samples, performance has not been improved accordingly. The quality, quantity, and diversity of the infrared dataset are critical to the detection of small targets. To highlight this issue, we propose a negative sample augmentation method in this paper. Specifically, a negative augmentation approach is proposed to generate massive negatives for self-supervised learning. Firstly, we perform a sequential noise modeling technology to generate realistic infrared data. Secondly, we fuse the extracted noise with the original data to facilitate diversity and fidelity in the generated data. Lastly, we proposed a negative augmentation strategy to enrich diversity as well as maintain semantic invariance. The proposed algorithm produces a synthetic SIRST-5K dataset, which contains massive pseudo-data and corresponding labels. With a rich diversity of infrared small target data, our algorithm significantly improves the model performance and convergence speed. Compared with other state-of-the-art (SOTA) methods, our method achieves outstanding performance in terms of probability of detection (Pd), false-alarm rate (Fa), and intersection over union (IoU). The code is available at: https://github.com/luy0222/SIRST-5K.

Recently, researchers have proposed various deep learning methods to accurately detect infrared targets with the characteristics of indistinct shape and texture. Due to the limited variety of infrared datasets, training deep learning models with good generalization poses a challenge. To augment the infrared dataset, researchers employ data augmentation techniques, which often involve generating new images by combining images from different datasets. However, these methods are lacking in two respects. In terms of realism, the images generated by mixup-based methods lack realism and are difficult to effectively simulate complex real-world scenarios. In terms of diversity, compared with real-world scenes, borrowing knowledge from another dataset inherently has a limited diversity. Currently, the diffusion model stands out as an innovative generative approach. Large-scale trained diffusion models have a strong generative prior that enables real-world modeling of images to generate diverse and realistic images. In this paper, we propose Diff-Mosaic, a data augmentation method based on the diffusion model. This model effectively alleviates the challenge of diversity and realism of data augmentation methods via diffusion prior. Specifically, our method consists of two stages. Firstly, we introduce an enhancement network called Pixel-Prior, which generates highly coordinated and realistic Mosaic images by harmonizing pixels. In the second stage, we propose an image enhancement strategy named Diff-Prior. This strategy utilizes diffusion priors to model images in the real-world scene, further enhancing the diversity and realism of the images. Extensive experiments have demonstrated that our approach significantly improves the performance of the detection network. The code is available at https://github.com/YupeiLin2388/Diff-Mosaic.

Deep learning technologies have demonstrated their effectiveness in removing cloud cover from optical remote-sensing images. Convolutional Neural Networks (CNNs) exert dominance in the cloud removal tasks. However, constrained by the inherent limitations of convolutional operations, CNNs can address only a modest fraction of cloud occlusion. In recent years, diffusion models have achieved state-of-the-art (SOTA) proficiency in image generation and reconstruction due to their formidable generative capabilities. Inspired by the rapid development of diffusion models, we first present an iterative diffusion process for cloud removal (IDF-CR), which exhibits a strong generative capabilities to achieve component divide-and-conquer cloud removal. IDF-CR consists of a pixel space cloud removal module (Pixel-CR) and a latent space iterative noise diffusion network (IND). Specifically, IDF-CR is divided into two-stage models that address pixel space and latent space. The two-stage model facilitates a strategic transition from preliminary cloud reduction to meticulous detail refinement. In the pixel space stage, Pixel-CR initiates the processing of cloudy images, yielding a suboptimal cloud removal prior to providing the diffusion model with prior cloud removal knowledge. In the latent space stage, the diffusion model transforms low-quality cloud removal into high-quality clean output. We refine the Stable Diffusion by implementing ControlNet. In addition, an unsupervised iterative noise refinement (INR) module is introduced for diffusion model to optimize the distribution of the predicted noise, thereby enhancing advanced detail recovery. Our model performs best with other SOTA methods, including image reconstruction and optical remote-sensing cloud removal on the optical remote-sensing datasets.

Modern visual recognition models often display overconfidence due to their reliance on complex deep neural networks and one-hot target supervision, resulting in unreliable confidence scores that necessitate calibration. While current confidence calibration techniques primarily address single-label scenarios, there is a lack of focus on more practical and generalizable multi-label contexts. This paper introduces the Multi-Label Confidence Calibration (MLCC) task, aiming to provide well-calibrated confidence scores in multi-label scenarios. Unlike single-label images, multi-label images contain multiple objects, leading to semantic confusion and further unreliability in confidence scores. Existing single-label calibration methods, based on label smoothing, fail to account for category correlations, which are crucial for addressing semantic confusion, thereby yielding sub-optimal performance. To overcome these limitations, we propose the Dynamic Correlation Learning and Regularization (DCLR) algorithm, which leverages multi-grained semantic correlations to better model semantic confusion for adaptive regularization. DCLR learns dynamic instance-level and prototype-level similarities specific to each category, using these to measure semantic correlations across different categories. With this understanding, we construct adaptive label vectors that assign higher values to categories with strong correlations, thereby facilitating more effective regularization. We establish an evaluation benchmark, re-implementing several advanced confidence calibration algorithms and applying them to leading multi-label recognition (MLR) models for fair comparison. Through extensive experiments, we demonstrate the superior performance of DCLR over existing methods in providing reliable confidence scores in multi-label scenarios.

Due to the advancement of deep learning, the performance of salient object detection (SOD) has been significantly improved. However, deep learning-based techniques require a sizable amount of pixel-wise annotations. To relieve the burden of data annotation, a variety of deep weakly-supervised and unsupervised SOD methods have been proposed, yet the performance gap between them and fully supervised methods remains significant. In this paper, we propose a novel, cost-efficient salient object detection framework, which can adapt models from synthetic data to real-world data with the help of a limited number of actively selected annotations. Specifically, we first construct a synthetic SOD dataset by copying and pasting foreground objects into pure background images. With the masks of foreground objects taken as the ground-truth saliency maps, this dataset can be used for training the SOD model initially. However, due to the large domain gap between synthetic images and real-world images, the performance of the initially trained model on the real-world images is deficient. To transfer the model from the synthetic dataset to the real-world datasets, we further design an uncertainty-aware active domain adaptive algorithm to generate labels for the real-world target images. The prediction variances against data augmentations are utilized to calculate the superpixel-level uncertainty values. For those superpixels with relatively low uncertainty, we directly generate pseudo labels according to the network predictions. Meanwhile, we select a few superpixels with high uncertainty scores and assign labels to them manually. This labeling strategy is capable of generating high-quality labels without incurring too much annotation cost. Experimental results on six benchmark SOD datasets demonstrate that our method outperforms the existing state-of-the-art weakly-supervised and unsupervised SOD methods and is even comparable to the fully supervised ones. Code will be released at: https://github.com/czh-3/UADA.

Video scene graph generation (VidSGG) aims to identify objects in visual scenes and infer their relationships for a given video. It requires not only a comprehensive understanding of each object scattered on the whole scene but also a deep dive into their temporal motions and interactions. Inherently, object pairs and their relationships enjoy spatial co-occurrence correlations within each image and temporal consistency/transition correlations across different images, which can serve as prior knowledge to facilitate VidSGG model learning and inference. In this work, we propose a spatial-temporal knowledge-embedded transformer (STKET) that incorporates the prior spatial-temporal knowledge into the multi-head cross-attention mechanism to learn more representative relationship representations. Specifically, we first learn spatial co-occurrence and temporal transition correlations in a statistical manner. Then, we design spatial and temporal knowledge-embedded layers that introduce the multi-head cross-attention mechanism to fully explore the interaction between visual representation and the knowledge to generate spatial- and temporal-embedded representations, respectively. Finally, we aggregate these representations for each subject-object pair to predict the final semantic labels and their relationships. Extensive experiments show that STKET outperforms current competing algorithms by a large margin, e.g., improving the mR@50 by 8.1%, 4.7%, and 2.1% on different settings over current algorithms.

Natural Language Video Localization (NLVL) has recently attracted much attention because of its practical significance. However, the existing methods still face the following challenges: 1) When the models learn intra-modal semantic association, the temporal causal interaction information and contextual semantic discriminative information are ignored, resulting in the lack of intra-modal semantic context connection; 2) When learning fusion representations, existing cross-modal interaction modules lack hierarchical attention function to extract inter-modal similarity information and intra-modal self-correlation information, resulting in insufficient cross-modal information interaction; and 3) When the loss function is optimized, the existing models ignore the correlation of causal inference between the start and end boundaries, resulting in inaccurate start and end boundary calibrations. To conquer the above challenges, we proposed a novel NLVL model, called Discriminative Parallel and Hierarchical Attention Network (DPHANet). Specifically, we emphasized the importance of temporal causal interaction information and contextual semantic discriminative information and correspondingly proposed a Discriminative Parallel Attention Encoder (DPAE) module to infer and encode the above critical information. Besides, to overcome the shortcomings of the existing cross-modal interaction modules, we designed a Video-Query Hierarchical Attention (VQHA) module, which can perform cross-modal interaction and intra-modal self-correlation modeling in a hierarchical manner. Furthermore, a novel deviation loss function was proposed to capture the correlation of causal inference between the start and end boundaries and force the model to focus on the continuity and temporal causality in the video. Finally, extensive experiments on three benchmark datasets demonstrated the superiority of our proposed DPHANet model, which has achieved about 1.5% and 3.5% average performance improvement and about 2.5% and 7.5% maximum performance improvement on the Charades-STA and TACoS datasets respectively.

The performance of existing methods for multi-person 3D pose estimation in crowded scenes is still limited, due to the challenge of heavy overlapping among persons. Attempt to address this issue, we propose a progressive inference scheme, i.e., Articulation-aware Knowledge Exploration (AKE), to improve the multi-person 3D pose models on those samples with complex occlusions at the inference stage. We argue it is beneficial to explore the underlying articulated information/knowledge of the human body, which helps to further correct the predicted poses in those samples. To exploit such information, we propose an iterative scheme to achieve a self-improving loop for keypoint association. Specifically, we introduce a kinematic validation module for locating unreasonable articulations and an occluded-keypoint discovering module for discovering occluded articulations. Extensive experiments on two challenging benchmarks under both weakly-supervised and fully-supervised settings demonstrate the superiority and generalization ability of our proposed method for crowded scenes.

Domain shift poses a significant challenge in Cross-Domain Facial Expression Recognition (CD-FER) due to the distribution variation between the source and target domains. Current algorithms mainly focus on learning domain-invariant features through global feature adaptation, while neglecting the transferability of local features across different domains. Additionally, these algorithms lack discriminative supervision during training on target datasets, resulting in deteriorated feature representation in the target domain. To address these limitations, we propose an Adaptive Global-Local Representation Learning and Selection (AGLRLS) framework. The framework incorporates global-local adversarial adaptation and semantic-aware pseudo label generation to enhance the learning of domain-invariant and discriminative feature representation during training. Meanwhile, a global-local prediction consistency learning is introduced to improve classification results during inference. Specifically, the framework consists of separate global-local adversarial learning modules that learn domain-invariant global and local features independently. We also design a semantic-aware pseudo label generation module, which computes semantic labels based on global and local features. Moreover, a novel dynamic threshold strategy is employed to learn the optimal thresholds by leveraging independent prediction of global and local features, ensuring filtering out the unreliable pseudo labels while retaining reliable ones. These labels are utilized for model optimization through the adversarial learning process in an end-to-end manner. During inference, a global-local prediction consistency module is developed to automatically learn an optimal result from multiple predictions. To validate the effectiveness of our framework, we conduct comprehensive experiments and analysis based on a fair evaluation benchmark. The results demonstrate that the proposed framework outperforms the current competing methods by a substantial margin.

As a cost-effective alternative to standard multi-label learning, the multi-label image recognition with partial positive labels (MLR-PPL) task attracts increasing attention, in which merely a portion of positive labels are given while the rest of positive labels and all negative labels are missing. To facilitate this task, we propose a novel framework that leverages semantic correlation among different images in a category-adaptive manner to complement unknown labels accurately. Specifically, the proposed framework consists of two complementary modules. 1) A category-adaptive label discovery (CALD) module is designed to measure the semantic similarity between positive samples and then complement unknown labels with high similarities. 2) A category-adaptive noise rejection (CANR) module is designed to compute the sample weights based on semantic similarities from different samples and discard noisy labels with low weights. Due to the various degrees of confidence calibration among different categories, searching appropriate thresholds for each category in the proposed framework is highly time-consuming. To avoid such a resource-intensive manual tuning, we introduce a category-adaptive threshold updating algorithm that introduces the category-specific positive and negative similarity to adjust the threshold adaptively. Extensive experiments on various benchmarks show that the proposed framework performs better than current state-of-the-art algorithms.

In real-world environments, outdoor imaging systems are often affected by disturbances such as rain degradation. Especially, in nighttime driving scenes, insufficient and uneven lighting shrouds the scenes in darkness, resulting degradation of both the image quality and visibility. Particularly, in the field of autonomous driving, the visual perception ability of RGB sensors experiences a sharp decline in such harsh scenarios. Additionally, driving assistance systems suffer from reduced capabilities in capturing and discerning the surrounding environment, posing a threat to driving safety. Single-view information captured by single-modal sensors cannot comprehensively depict the entire scene. To address these challenges, we developed an image deraining framework tailored for rainy nighttime driving scenes. It aims to remove rain artifacts, enrich scene representation, and restore useful information. Specifically, we introduce cooperative learning between visible and infrared images captured by different sensors. By cross-view fusion of these multi-source data, the scene within the images gains richer texture details and enhanced contrast. We constructed an information cleaning module called CleanNet as the first stage of our framework. Moreover, we designed an information fusion module called FusionNet as the second stage to fuse the clean visible images with infrared images. Using this stage-by-stage learning strategy, we obtain de-rained fusion images with higher quality and better visual perception. Extensive experiments demonstrate the effectiveness of our proposed Cross-View Cooperative Learning (CVCL) in adverse driving scenarios in low-light rainy environments. The proposed approach addresses the gap in the utilization of existing rain removal algorithms in specific low-light conditions. It also holds promise for extending the application of image de-raining and image fusion methods to computer vision tasks.

Along with the concern of similarity measures in linear space, supervised online hash methods have been applied to the retrieval task. However, they ignored multi-dimensional space semantic mining and association characteristics will cause quantization errors of information hash code: 1) The similarity relation of discretized data needs to be considered in different spaces; 2) Latent semantic features need to be continuously embedded into hash code learning; 3) The correlation between the structure similarity and discrete hash matrices needs to be continuously optimized. To tackle these challenges, this paper proposes a novel Extensible Max-min Collaborative Retention Online Hash retrieval method based on mini-batch training data (EMCROH). It mainly includes the Max-min Bayesian Similarity Sparse Latent Hash module (MBSSLH), and the Repetition Collaborative Projection Learning module (RCPL). Specifically, MBSSLH is a max-min optimization model. Firstly, to explore the semantic similarity of multi-dimensional space, we propose a novel liner and nonlinear semantic similarity discrimination mechanism based on the log maximum likelihood similarity estimation with Euclidean space and minimize the input batch data features with a common projection matrix. Moreover, to further mine the potential semantic information of the discretization, we also propose a robust sparse discrete latent semantic information extraction submodule based on double latent factors. RCPL can extend the data externally using the repetition collaborative projection matrix with robustness regularization constraint. Finally, a novel max-min embedding iterative step is proposed to solve the batch discrete optimization problem based on Augmented Lagrange Multipliers (ALM) with Alternating Direction Minimization (ADM). Extensive experiments on several well-known large databases demonstrate that EMCROH outperforms the state-of-the-art hash methods. Code and datasets have been publicly available athttps://github.com/Tjeep-Tan/EMCROH.

While deep neural networks have achieved remarkable performance, they tend to lack transparency in prediction. The pursuit of greater interpretability in neural networks often results in a degradation of their original performance. Some works strive to improve both interpretability and performance, but they primarily depend on meticulously imposed conditions. In this paper, we propose a simple yet effective framework that acquires more explainable activation heatmaps and simultaneously increases the model performance, without the need for any extra supervision. Specifically, our concise framework introduces a new metric, i.e., explanation consistency, to reweight the training samples adaptively in model learning. The explanation consistency metric is utilized to measure the similarity between the model's visual explanations of the original samples and those of semantic-preserved adversarial samples, whose background regions are perturbed by using image adversarial attack techniques. Our framework then promotes the model learning by paying closer attention to those training samples with a high difference in explanations (i.e., low explanation consistency), for which the current model cannot provide robust interpretations. Comprehensive experimental results on various benchmarks demonstrate the superiority of our framework in multiple aspects, including higher recognition accuracy, greater data debiasing capability, stronger network robustness, and more precise localization ability on both regular networks and interpretable networks. We also provide extensive ablation studies and qualitative analyses to unveil the detailed contribution of each component.

Learning to recognize novel concepts from just a few image samples is very challenging as the learned model is easily overfitted on the few data and results in poor generalizability. One promising but underexplored solution is to compensate for the novel classes by generating plausible samples. However, most existing works of this line exploit visual information only, rendering the generated data easy to be distracted by some challenging factors contained in the few available samples. Being aware of the semantic information in the textual modality that reflects human concepts, this work proposes a novel framework that exploits semantic relations to guide dual-view data hallucination for few-shot image recognition. The proposed framework enables generating more diverse and reasonable data samples for novel classes through effective information transfer from base classes. Specifically, an instance-view data hallucination module hallucinates each sample of a novel class to generate new data by employing local semantic correlated attention and global semantic feature fusion derived from base classes. Meanwhile, a prototype-view data hallucination module exploits semantic-aware measure to estimate the prototype of a novel class and the associated distribution from the few samples, which thereby harvests the prototype as a more stable sample and enables resampling a large number of samples. We conduct extensive experiments and comparisons with state-of-the-art methods on several popular few-shot benchmarks to verify the effectiveness of the proposed framework.

Heavy rain significantly reduces image visibility, hindering tasks like autonomous driving and video surveillance. Many existing rain removal methods, while effective in light rain, falter under heavy rain due to their reliance on purely spatial features. Recognizing this challenge, we introduce the Wavelet-Spatial Dual Attention Transformer Framework (WSDformer). This innovative architecture adeptly captures both frequency and spatial characteristics, anchored by the wavelet-spatial dual attention (WSDA) mechanism. While the spatial attention zeroes in on intricate local details, the wavelet attention leverages wavelet decomposition to encompass diverse frequency information, augmenting the spatial representations. Furthermore, addressing the persistent issue of incomplete structural detail restoration, we integrate the PriorFormer Block (PFB). This unique module, underpinned by the Prior Fusion Attention (PFA), synergizes residual channel prior features with input features, thereby enhancing background structures and guiding precise rain feature extraction. To navigate the intrinsic constraints of U-shaped transformers, such as semantic discontinuities and subdued multi-scale interactions from skip connections, our Cross Interaction U-Shaped Transformer Network is introduced. This design empowers superior semantic layers to streamline the extraction of their lower-tier counterparts, optimizing network learning. Empirical analysis reveals our method's leading prowess across rainy image datasets and achieves state-of-the-art performance, with notable supremacy in heavy rainfall conditions. This superiority extends to diverse visual challenges and real-world rainy scenarios, affirming its broad applicability and robustness. The source code is available athttps://github.com/Jiongze-Yu/WSDformer.

Graph-level representations are critical in various real-world applications, such as predicting the properties of molecules. However, in practice, precise graph annotations are generally very expensive and time-consuming. To address this issue, graph contrastive learning constructs an instance discrimination task, which pulls together positive pairs (augmentation pairs of the same graph) and pushes away negative pairs (augmentation pairs of different graphs) for unsupervised representation learning. However, since for a query, its negatives are uniformly sampled from all graphs, existing methods suffer from the critical sampling bias issue, i.e., the negatives likely having the same semantic structure with the query, leading to performance degradation. To mitigate this sampling bias issue, in this article, we propose a prototypical graph contrastive learning (PGCL) approach. Specifically, PGCL models the underlying semantic structure of the graph data via clustering semantically similar graphs into the same group and simultaneously encourages the clustering consistency for different augmentations of the same graph. Then, given a query, it performs negative sampling via drawing the graphs from those clusters that differ from the cluster of query, which ensures the semantic difference between query and its negative samples. Moreover, for a query, PGCL further reweights its negative samples based on the distance between their prototypes (cluster centroids) and the query prototype such that those negatives having moderate prototype distance enjoy relatively large weights. This reweighting strategy is proven to be more effective than uniform sampling. Experimental results on various graph benchmarks testify the advantages of our PGCL over state-of-the-art methods. The code is publicly available at https://github.com/ha-lins/PGCL.

Since math word problem (MWP) solving aims to transform natural language problem description into executable solution equations, an MWP solver needs to not only comprehend the real-world narrative described in the problem text but also identify the relationships among the quantifiers and variables implied in the problem and maps them into a reasonable solution equation logic. Recently, although deep learning models have made great progress in MWPs, they ignore the grounding equation logic implied by the problem text. Besides, as we all know, pretrained language models (PLM) have a wealth of knowledge and high-quality semantic representations, which may help solve MWPs, but they have not been explored in the MWP-solving task. To harvest the equation logic and real-world knowledge, we propose a template-based contrastive distillation pretraining (TCDP) approach based on a PLM-based encoder to incorporate mathematical logic knowledge by multiview contrastive learning while retaining rich real-world knowledge and high-quality semantic representation via knowledge distillation. We named the pretrained PLM-based encoder by our approach as MathEncoder. Specifically, the mathematical logic is first summarized by clustering the symbolic solution templates among MWPs and then injected into the deployed PLM-based encoder by conducting supervised contrastive learning based on the symbolic solution templates, which can represent the underlying solving logic in the problems. Meanwhile, the rich knowledge and high-quality semantic representation are retained by distilling them from a well-trained PLM-based teacher encoder into our MathEncoder. To validate the effectiveness of our pretrained MathEncoder, we construct a new solver named MathSolver by replacing the GRU-based encoder with our pretrained MathEncoder in GTS, which is a state-of-the-art MWP solver. The experimental results demonstrate that our method can carry a solver's understanding ability of MWPs to a new stage by outperforming existing state-of-the-art methods on two widely adopted benchmarks Math23K and CM17K. Code will be available at https://github.com/QinJinghui/tcdp.

Popularity bias, as a long-standing problem in recommender systems (RSs), has been fully considered and explored for offline recommendation systems in most existing relevant researches, but very few studies have paid attention to eliminate such bias in online interactive recommendation scenarios. Bias amplification will become increasingly serious over time due to the existence of feedback loop between the user and the interactive system. However, existing methods have only investigated the causal relations among different factors statically without considering temporal dependencies inherent in the online interactive recommendation system, making them difficult to be adapted to online settings. To address these problems, we propose a novel counterfactual interactive policy learning (CIPL) method to eliminate popularity bias for online recommendation. It first scrutinizes the causal relations in the interactive recommender models and formulates a novel temporal causal graph (TCG) to guide the training and counterfactual inference of the causal interactive recommendation system. Concretely, TCG is used to estimate the causal relations of item popularity on prediction score when the user interacts with the system at each time during model training. Besides, it is also used to remove the negative effect of popularity bias in the test stage. To train the causal interactive recommendation system, we formulated our CIPL by the actor-critic framework with an online interactive environment simulator. We conduct extensive experiments on three public benchmarks and the experimental results demonstrate that our proposed method can achieve the new state-of-the-art performance.

To facilitate more accurate and explainable recommendation, it is crucial to incorporate side information into user-item interactions. Recently, knowledge graph (KG) has attracted much attention in a variety of domains due to its fruitful facts and abundant relations. However, the expanding scale of real-world data graphs poses severe challenges. In general, most existing KG-based algorithms adopt exhaustively hop-by-hop enumeration strategy to search all the possible relational paths, this manner involves extremely high-cost computations and is not scalable with the increase of hop numbers. To overcome these difficulties, in this article, we propose an end-to-end framework Knowledge-tree-routed UseR-Interest Trajectories Network (KURIT-Net). KURIT-Net employs the user-interest Markov trees (UIMTs) to reconfigure a recommendation-based KG, striking a good balance for routing knowledge between short-distance and long-distance relations between entities. Each tree starts from the preferred items for a user and routes the association reasoning paths along the entities in the KG to provide a human-readable explanation for model prediction. KURIT-Net receives entity and relation trajectory embedding (RTE) and fully reflects potential interests of each user by summarizing all reasoning paths in a KG. Besides, we conduct extensive experiments on six public datasets, our KURIT-Net significantly outperforms state-of-the-art approaches and shows its interpretability in recommendation.

Text-driven 3D scene editing has gained significant attention owing to its convenience and user-friendliness. However, existing methods still lack accurate control of the specified appearance and location of the editing result due to the inherent limitations of the text description. To this end, we propose a 3D scene editing framework, TIP-Editor, that accepts both text and image prompts and a 3D bounding box to specify the editing region. With the image prompt, users can conveniently specify the detailed appearance/style of the target content in complement to the text description, enabling accurate control of the appearance. Specifically, TIP-Editor employs a stepwise 2D personalization strategy to better learn the representation of the existing scene and the reference image, in which a localization loss is proposed to encourage correct object placement as specified by the bounding box. Additionally, TIP-Editor utilizes explicit and flexible 3D Gaussian splatting (GS) as the 3D representation to facilitate local editing while keeping the background unchanged. Extensive experiments have demonstrated that TIP-Editor conducts accurate editing following the text and image prompts in the specified bounding box region, consistently outperforming the baselines in editing quality, and the alignment to the prompts, qualitatively and quantitatively.

Vision-Language Navigation (VLN) is a challenging task which requires an agent to align complex visual observations to language instructions to reach the goal position. Most existing VLN agents directly learn to align the raw directional features and visual features trained using one-hot labels to linguistic instruction features. However, the big semantic gap among these multi-modal inputs makes the alignment difficult and therefore limits the navigation performance. In this paper, we propose Actional Atomic-Concept Learning (AACL), which maps visual observations to actional atomic concepts for facilitating the alignment. Specifically, an actional atomic concept is a natural language phrase containing an atomic action and an object, e.g., ``go up stairs''. These actional atomic concepts, which serve as the bridge between observations and instructions, can effectively mitigate the semantic gap and simplify the alignment. AACL contains three core components: 1) a concept mapping module to map the observations to the actional atomic concept representations through the VLN environment and the recently proposed Contrastive Language-Image Pretraining (CLIP) model, 2) a concept refining adapter to encourage more instruction-oriented object concept extraction by re-ranking the predicted object concepts by CLIP, and 3) an observation co-embedding module which utilizes concept representations to regularize the observation representations. Our AACL establishes new state-of-the-art results on both fine-grained (R2R) and high-level (REVERIE and R2R-Last) VLN benchmarks. Moreover, the visualization shows that AACL significantly improves the interpretability in action decision. Code will be available at https://gitee.com/mindspore/models/tree/master/research/cv/VLN-AACL.

Semi-supervised object detection (SSOD) attracts extensive research interest due to its great significance in reducing the data annotation effort. Collecting high-quality and category-balanced pseudo labels for unlabeled images is critical to addressing the SSOD problem. However, most of the existing pseudo-labeling-based methods depend on a large and fixed threshold to select high-quality pseudo labels from the predictions of a teacher model. Considering different object classes usually have different detection difficulty levels due to scale variance and data distribution imbalance, conventional pseudo-labeling-based methods are arduous to explore the value of unlabeled data sufficiently. To address these issues, we propose an adaptive pseudo labeling strategy, which can assign thresholds to classes with respect to their “hardness”. This is beneficial for ensuring the high quality of easier classes and increasing the quantity of harder classes simultaneously. Besides, label refinement modules are set up based on box jittering for guaranteeing the localization quality of pseudo labels. To further improve the algorithm’s robustness against scale variance and make the most of pseudo labels, we devise a joint feature-level and prediction-level consistency learning pipeline for transferring the information of the teacher model to the student model. Extensive experiments on COCO and VOC datasets indicate that our method achieves state-of-the-art performance. Especially, it brings mean average precision gains of 2.08 and 1.28 on MS-COCO dataset with 5% and 10% labeled images, respectively.

Most of the recent research in semi-supervised object detection follows the pseudo-labeling paradigm evolved from the semi-supervised image classification task. However, the training paradigm of the two-stage object detector inevitably makes the pseudo-label learning process for unlabeled images full of bias. Specifically, the IoU matching scheme used for selecting and labeling candidate boxes is based on the assumption that the matching source~(ground truth) is accurate enough in terms of the number of objects, object position and object category. Obviously, pseudo-labels generated for unlabeled images cannot satisfy such a strong assumption, which makes the produced training proposals extremely unreliable and thus severely spoil the follow-up training. To de-bias the training proposals generated by the pseudo-label-based IoU matching, we propose a general framework -- De-biased Teacher, which abandons both the IoU matching and pseudo labeling processes by directly generating favorable training proposals for consistency regularization between the weak/strong augmented image pairs. Moreover, a distribution-based refinement scheme is designed to eliminate the scattered class predictions of significantly low values for higher efficiency. Extensive experiments demonstrate that the proposed De-biased Teacher consistently outperforms other state-of-the-art methods on the MS-COCO and PASCAL VOC benchmarks. Source codes are available at https://github.com/wkfdb/De-biased-Teracher.

In this paper, we study graph-to-image generation conditioned exclusively on scene graphs, in which we seek to disentangle the veiled semantics between knowledge graphs and images. While most existing research resorts to laborious auxiliary information such as object layouts or segmentation masks, it is also of interest to unveil the generality of the model with limited supervision, moreover, avoiding extra cross-modal alignments. To tackle this challenge, we delve into the causality of the adversarial generation process, and reason out a new principle to realize a simultaneous semantic disentanglement with an alignment on target and model distributions. This principle is named knowledge consensus, which explicitly describes a triangle causal dependency among observed images, graph semantics and hidden visual representations. The consensus also determines a new graph-to-image generation framework, carried on several adversarial optimization objectives. Extensive experimental results demonstrate that, even conditioned only on scene graphs, our model surprisingly achieves superior performance on semantics-aware image generation, without losing the competence on manipulating the generation through knowledge graphs.

Text-to-motion generation is an emerging and challenging problem, which aims to synthesize motion with the same semantics as the input text. However, due to the lack of diverse labeled training data, most approaches either limit to specific types of text annotations or require online optimizations to cater to the texts during inference at the cost of efficiency and stability. In this paper, we investigate offline open-vocabulary text-to-motion generation in a zero-shot learning manner that neither requires paired training data nor extra online optimization to adapt for unseen texts. Inspired by the prompt learning in NLP, we pretrain a motion generator that learns to reconstruct the full motion from the masked motion. During inference, instead of changing the motion generator, our method reformulates the input text into a masked motion as the prompt for the motion generator to “reconstruct” the motion. In constructing the prompt, the unmasked poses of the prompt are synthesized by a text-to-pose generator. To supervise the optimization of the text-to-pose generator, we propose the first text-pose alignment model for measuring the alignment between texts and 3D poses. And to prevent the pose generator from over-fitting to limited training texts, we further propose a novel wordless training mechanism that optimizes the text-to-pose generator without any training texts. The comprehensive experimental results show that our method obtains a significant improvement against the baseline methods. The code is available at https://github.com/junfanlin/oohmg.

Deep learning models are challenged by the distribution shift between the training data and test data. Recently, the large models pre-trained on diverse data have demonstrated unprecedented robustness to various distribution shifts. However, fine-tuning these models can lead to a trade-off between in-distribution (ID) performance and out-of-distribution (OOD) robustness. Existing methods for tackling this trade-off do not explicitly address the OOD robustness problem. In this paper, based on causal analysis of the aforementioned problems, we propose a novel fine-tuning method, which uses masked images as counterfactual samples that help improve the robustness of the fine-tuning model. Specifically, we mask either the semantics-related or semantics-unrelated patches of the images based on class activation map to break the spurious correlation, and refill the masked patches with patches from other images. The resulting counterfactual samples are used in feature-based distillation with the pre-trained model. Extensive experiments verify that regularizing the fine-tuning with the proposed masked images can achieve a better trade-off between ID and OOD performance, surpassing previous methods on the OOD performance. Our code is available at https://github.com/Coxy7/robust-finetuning.

Generating talking face videos from audio attracts lots of research interest. A few person-specific methods can generate vivid videos but require the target speaker's videos for training or fine-tuning. Existing person-generic methods have difficulty in generating realistic and lip-synced videos while preserving identity information. To tackle this problem, we propose a two-stage framework consisting of audio-to-landmark generation and landmark-to-video rendering procedures. First, we devise a novel Transformer-based landmark generator to infer lip and jaw landmarks from the audio. Prior landmark characteristics of the speaker's face are employed to make the generated landmarks coincide with the facial outline of the speaker. Then, a video rendering model is built to translate the generated landmarks into face images. During this stage, prior appearance information is extracted from the lower-half occluded target face and static reference images, which helps generate realistic and identity-preserving visual content. For effectively exploring the prior information of static reference images, we align static reference images with the target face's pose and expression based on motion fields. Moreover, auditory features are reused to guarantee that the generated face images are well synchronized with the audio. Extensive experiments demonstrate that our method can produce more realistic, lip-synced, and identity-preserving videos than existing person-generic talking face generation methods.

Conversational Recommender System (CRS)aims to explicitly acquire user preferences towards items and attributes through natural language conversations.However, existing CRS methods ask users to provide explicit answers (yes/no) for each attribute they require, regardless of users' knowledge or interest, which may significantly reduce the user experience and semantic consistency.Furthermore, these methods assume that users like all attributes of the target item and dislike those unrelated to it, which can introduce bias in attributelevel feedback and impede the system's ability to accurately identify the target item.To address these issues, we propose a more realistic, user-friendly, and explainable CRS framework called Hierarchical User-Interest Tracking for Conversational Recommender System (HutCRS).HutCRS portrays the conversation as a hierarchical interest tree that consists of two stages.In stage I, the system identifies the aspects that the user prefers while the system asks about attributes related to these positive aspects or recommends items in stage II.In addition, we develop a Hierarchical-Interest Policy Learning (HIPL) module to integrate the decision-making process of which aspects to ask and when to ask about attributes or recommend items.Moreover, we classify the attribute-level feedback results to further enhance the system's ability to capture special information, such as attribute instances that are accepted by users but not presented in their historical interactive data.Extensive experiments on four benchmark datasets demonstrate the superiority of our method.The implementation of HutCRS is publicly available at https://github.com/xinle1129/HutCRS.

The vision community is undergoing the unprecedented progress with the emergence of Vision-Language Pretraining Models (VLMs). Prompt learning plays as the holy grail of accessing VLMs since it enables their fast adaptation to downstream tasks with limited resources. Whereas existing researches milling around single-prompt paradigms, rarely investigate the technical potential behind their multi-prompt learning counterparts. This paper aims to provide a principled retrospect for vision-language multi-prompt learning. We extend the recent constant modality gap phenomenon to learnable prompts and then, justify the superiority of vision-language transfer with multi-prompt augmentation, empirically and theoretically. In terms of this observation, we propose an Energy-based Multi-prompt Learning (EMPL) to generate multiple prompt embeddings by drawing instances from an energy-based distribution, which is implicitly defined by VLMs. So our EMPL is not only parameter-efficient but also rigorously lead to the balance between in-domain and out-of-domain open-vocabulary generalization. Comprehensive experiments have been conducted to justify our claims and the excellence of EMPL.

The self-attention mechanism (SAM) is widely used in various fields of artificial intelligence and has successfully boosted the performance of different models. However, current explanations of this mechanism are mainly based on intuitions and experiences, while there still lacks direct modeling for how the SAM helps performance. To mitigate this issue, in this paper, based on the dynamical system perspective of the residual neural network, we first show that the intrinsic stiffness phenomenon (SP) in the high-precision solution of ordinary differential equations (ODEs) also widely exists in high-performance neural networks (NN). Thus the ability of NN to measure SP at the feature level is necessary to obtain high performance and is an important factor in the difficulty of training NN. Similar to the adaptive step-size method which is effective in solving stiff ODEs, we show that the SAM is also a stiffness-aware step size adaptor that can enhance the model's representational ability to measure intrinsic SP by refining the estimation of stiffness information and generating adaptive attention values, which provides a new understanding about why and how the SAM can benefit the model performance. This novel perspective can also explain the lottery ticket hypothesis in SAM, design new quantitative metrics of representational ability, and inspire a new theoretic-inspired approach, StepNet. Extensive experiments on several popular benchmarks demonstrate that StepNet can extract fine-grained stiffness information and measure SP accurately, leading to significant improvements in various visual tasks.

Multi-person 3D mesh recovery from videos is a critical first step towards automatic perception of group behavior in virtual reality, physical therapy and beyond. However, existing approaches rely on multi-stage paradigms, where the person detection and tracking stages are performed in a multi-person setting, while temporal dynamics are only modeled for one person at a time. Consequently, their performance is severely limited by the lack of inter-person interactions in the spatial-temporal mesh recovery, as well as by detection and tracking defects. To address these challenges, we propose the Coordinate transFormer (Coord-Former) that directly models multi-person spatial-temporal relations and simultaneously performs multi-mesh recovery in an end-to-end manner Instead of partitioning the feature map into coarse-scale patch-wise tokens, CoordFormer leverages a novel Coordinate-Aware Attention to preserve pixel-level spatial-temporal coordinate information. Additionally, we propose a simple, yet effective Body Center Attention mechanism to fuse position information. Extensive experiments on the 3DPW dataset demonstrate that CoordFormer significantly improves the state-of-the-art, outperforming the previously best results by 4.2%, 8.8% and 4.7% according to the MPJPE, PAMPJPE, and PVE metrics, respectively, while being 40% faster than recent video-based approaches. The released code can be found at https://github.com/Li-Hao-yuan/CoordFormer.

As a mainstream framework in the field of semi-supervised learning (SSL), self-training via pseudo labeling and its variants have witnessed impressive progress in semi-supervised semantic segmentation with the recent advance of deep neural networks. However, modern self-training based SSL algorithms use a pre-defined constant threshold to select unlabeled pixel samples that contribute to the training, thus failing to be compatible with different learning difficulties of variant categories and different learning status of the model. To address these issues, we propose Enhanced Soft Label (ESL), a curriculum learning approach to fully leverage the high-value supervisory signals implicit in the untrustworthy pseudo label. ESL believes that pixels with unconfident predictions can be pretty sure about their belonging to a subset of dominant classes though being arduous to determine the exact one. It thus contains a Dynamic Soft Label (DSL) module to dynamically maintain the high probability classes, keeping the label "soft" so as to make full use of the high entropy prediction. However, the DSL itself will inevitably introduce ambiguity between dominant classes, thus blurring the classification boundary. Therefore, we further propose a pixel-to-part contrastive learning method cooperated with an unsupervised object part grouping mechanism to improve its ability to distinguish between different classes. Extensive experimental results on Pascal VOC 2012 and Cityscapes show that our approach achieves remarkable improvements over existing state-of-the-art approaches.

Despite substantial advances, single-image super-resolution (SISR) is always in a dilemma to reconstruct high-quality images with limited information from one input image, especially in realistic scenarios. In this paper, we establish a large-scale real-world burst super-resolution dataset, i.e., RealBSR, to explore the faithful reconstruction of image details from multiple frames. Furthermore, we introduce a Federated Burst Affinity network (FBAnet) to investigate non-trivial pixel-wise displacements among images under real-world image degradation. Specifically, rather than using pixel-wise alignment, our FBAnet employs a simple homography alignment from a structural geometry aspect and a Federated Affinity Fusion (FAF) strategy to aggregate the complementary information among frames. Those fused informative representations are fed to a Transformer-based module of burst representation decoding. Besides, we have conducted extensive experiments on two versions of our datasets, i.e., RealBSR-RAW and RealBSR-RGB. Experimental results demonstrate that our FBAnet outperforms existing state-of-the-art burst SR methods and also achieves visually-pleasant SR image predictions with model details. Our dataset, codes, and models are publicly available at https://github.com/yjsunnn/FBANet.

Skeleton sequence representation learning has shown great advantages for action recognition due to its promising ability to model human joints and topology. However, the current methods usually require sufficient labeled data for training computationally expensive models. Moreover, these methods ignore how to utilize the fine-grained dependencies among different skeleton joints to pre-train an efficient skeleton sequence learning model that can generalize well across different datasets. In this paper, we propose an efficient skeleton sequence learning framework, named Skeleton Sequence Learning (SSL). To comprehensively capture the human pose and obtain discriminative skeleton sequence representation, we build an asymmetric graph-based encoder-decoder pre-training architecture named SkeletonMAE, which embeds skeleton joint sequence into graph convolutional network and reconstructs the masked skeleton joints and edges based on the prior human topology knowledge. Then, the pre-trained SkeletonMAE encoder is integrated with the Spatial-Temporal Representation Learning (STRL) module to build the SSL framework. Extensive experimental results show that our SSL generalizes well across different datasets and outperforms the state-of-the-art self-supervised skeleton-based methods on FineGym, Diving48, NTU 60 and NTU 120 datasets. Moreover, we obtain comparable performance to some fully supervised methods. The code is avaliable at https://github.com/HongYan1123/SkeletonMAE.

Thanks to the rapid development of diffusion models, unprecedented progress has been witnessed in image synthesis. Prior works mostly rely on pre-trained linguistic models, but a text is often too abstract to properly specify all the spatial properties of an image, e.g., the layout configuration of a scene, leading to the sub-optimal results of complex scene generation. In this paper, we achieve accurate complex scene generation by proposing a semantically controllable Layout-AWare diffusion model, termed LAW-Diffusion. Distinct from the previous Layout-to-Image generation (L2I) methods that primarily explore category-aware relationships, LAW-Diffusion introduces a spatial dependency parser to encode the location-aware semantic coherence across objects as a layout embedding and produces a scene with perceptually harmonious object styles and contextual relations. To be specific, we delicately instantiate each object’s regional semantics as an object region map and leverage a location-aware cross-object attention module to capture the spatial dependencies among those disentangled representations. We further propose an adaptive guidance schedule for our layout guidance to mitigate the trade-off between the regional semantic alignment and the texture fidelity of generated objects. Moreover, LAW-Diffusion allows for instance reconfiguration while maintaining the other regions in a synthesized image by introducing a layout-aware latent grafting mechanism to recompose its local regional semantics. To better verify the plausibility of generated scenes, we propose a new evaluation metric for the L2I task, dubbed Scene Relation Score (SRS) to measure how the images preserve the rational and harmonious relations among contextual objects. Comprehensive experiments on COCO-Stuff and Visual-Genome demonstrate that our LAW-Diffusion yields the state-of-the-art generative performance, especially with coherent object relations.

Learning with noisy labels (LNL) is one of the most important and challenging problems in weakly-supervised learning. Recent advances adopt the sample selection strategy to mitigate the interference of noisy labels and use small-loss criteria to select clean samples. However, the one-dimensional loss is an over-simplified metric that fails to accommodate the complex feature landscape of various samples, and, hence, is prone to introduce classification errors during sample selection. In this paper, we propose RankMatch, a novel LNL framework that investigates additional dimensions of confidence and consistency in order to combat noisy labels. Confidence-wise, we propose a novel sample selection strategy based on confidence representation voting instead of the widely-used small-loss criterion. This new strategy is capable of increasing sample selection quantity without sacrificing labeling accuracy. Consistency-wise, instead of the widely adopted feature distance metric for measuring the consistency of inner-class samples, we advocate that the rank of principal features is a much more robust indicator. Based on this metric, we propose rank contrastive loss, which strengthens the consistency of similar samples regardless of their labels and facilitates feature representation learning. Experimental results on noisy versions of CIFAR-10, CIFAR-100, Clothing1M and WebVision have validated the superiority of our approach over existing state-of-the-art methods.

Cross-modal garment synthesis and manipulation will significantly benefit the way fashion designers generate garments and modify their designs via flexible linguistic interfaces. However, despite the significant progress that has been made in generic image synthesis using diffusion models, producing garment images with garment part level semantics that are well aligned with input text prompts and then flexibly manipulating the generated results still remains a problem. Current approaches follow the general text-to-image paradigm and mine cross-modal relations via simple cross-attention modules, neglecting the structural correspondence between visual and textual representations in the fashion design domain. In this work, we instead introduce DiffCloth, a diffusion-based pipeline for cross-modal garment synthesis and manipulation, which empowers diffusion models with flexible compositionality in the fashion domain by structurally aligning the cross-modal semantics. Specifically, we formulate the part-level cross-modal alignment as a bipartite matching problem between the linguistic Attribute-Phrases (AP) and the visual garment parts which are obtained via constituency parsing and semantic segmentation, respectively. To mitigate the issue of attribute confusion, we further propose a semantic-bundled cross-attention to preserve the spatial structure similarities between the attention maps of attribute adjectives and part nouns in each AP. Moreover, DiffCloth allows for manipulation of the generated results by simply replacing APs in the text prompts. The manipulation-irrelevant regions are recognized by blended masks obtained from the bundled attention maps of the APs and kept unchanged. Extensive experiments on the CM-Fashion benchmark demonstrate that DiffCloth both yields state-of-the-art garment synthesis results by leveraging the inherent structural information and supports flexible manipulation with region consistency.

Unsupervised Domain Adaptation (UDA) learns a model for an unlabeled target domain, utilizing a labeled source domain. Most existing works on UDA assume the availability of source data and neglect the adversarial robustness of the models, hindering security-sensitive real-world applications. In this paper, we study adversarially robust source-free UDA, aiming to train a robust target model by adapting a non-robust source model without using source data. A basic approach is to train a non-robust teacher model via conventional source-free UDA to predict pseudo-labels for target data, and then train a robust student model via adversarial training (AT). However, AT tends to magnify the errors of the teacher model, reducing the accuracy. Hence, we propose Relaxed Adversarial Training (RAT) that relieves the constraints on the confidence of predictions in AT to balance the robustness and accuracy. Extensive experiments validate that RAT can improve the accuracy on clean and adversarial samples, and is superior to related methods. Our code is available at https://github.com/Coxy7/RAT.

Traffic Signal Control (TSC) aims to reduce the average travel time of vehicles in a road network, which in turn enhances fuel utilization efficiency, air quality, and road safety, benefiting society as a whole. Due to the complexity of long-horizon control and coordination, most prior TSC methods leverage deep reinforcement learning (RL) to search for a control policy and have witnessed great success. However, TSC still faces two significant challenges. 1) The travel time of a vehicle is delayed feedback on the effectiveness of TSC policy at each traffic intersection since it is obtained after the vehicle has left the road network. Although several heuristic reward functions have been proposed as substitutes for travel time, they are usually biased and not leading the policy to improve in the correct direction. 2) The traffic condition of each intersection is influenced by the non-local intersections since vehicles traverse multiple intersections over time. Therefore, the TSC agent is required to leverage both the local observation and the non-local traffic conditions to predict the long-horizontal traffic conditions of each intersection comprehensively. To address these challenges, we propose DenseLight, a novel RL-based TSC method that employs an unbiased reward function to provide dense feedback on policy effectiveness and a non-local enhanced TSC agent to better predict future traffic conditions for more precise traffic control. Extensive experiments and ablation studies demonstrate that DenseLight can consistently outperform advanced baselines on various road networks with diverse traffic flows. The code is available at https://github.com/junfanlin/DenseLight.

Wind power is attracting increasing attention around the world due to its renewable, pollution-free, and other advantages. However, safely and stably integrating the high permeability intermittent power energy into electric power systems remains challenging. Accurate wind power forecasting (WPF) can effectively reduce power fluctuations in power system operations. Existing methods are mainly designed for short-term predictions and lack effective spatial-temporal feature augmentation. In this work, we propose a novel end-to-end wind power forecasting model named Hierarchical Spatial-Temporal Transformer Network (HSTTN) to address the long-term WPF problems. Specifically, we construct an hourglass-shaped encoder-decoder framework with skip-connections to jointly model representations aggregated in hierarchical temporal scales, which benefits long-term forecasting. Based on this framework, we capture the inter-scale long-range temporal dependencies and global spatial correlations with two parallel Transformer skeletons and strengthen the intra-scale connections with downsampling and upsampling operations. Moreover, the complementary information from spatial and temporal features is fused and propagated in each other via Contextual Fusion Blocks (CFBs) to promote the prediction further. Extensive experimental results on two large-scale real-world datasets demonstrate the superior performance of our HSTTN over existing solutions.

Existing methods for video question answering (VideoQA) often suffer from spurious correlations between different modalities, leading to a failure in identifying the dominant visual evidence and the intended question. Moreover, these methods function as black boxes, making it difficult to interpret the visual scene during the QA process. In this paper, to discover critical video segments and frames that serve as the visual causal scene for generating reliable answers, we present a causal analysis of VideoQA and propose a framework for cross-modal causal relational reasoning, named Visual Causal Scene Refinement (VCSR). Particularly, a set of causal front-door intervention operations is introduced to explicitly find the visual causal scenes at both segment and frame levels. Our VCSR involves two essential modules: i) the Question-Guided Refiner (QGR) module, which refines consecutive video frames guided by the question semantics to obtain more representative segment features for causal front-door intervention; ii) the Causal Scene Separator (CSS) module, which discovers a collection of visual causal and non-causal scenes based on the visual-linguistic causal relevance and estimates the causal effect of the scene-separating intervention in a contrastive learning manner. Extensive experiments on the NExT-QA, Causal-VidQA, and MSRVTT-QA datasets demonstrate the superiority of our VCSR in discovering visual causal scene and achieving robust video question answering.

Diffusion models, which have emerged to become popular text-to-image generation models, can produce high-quality and content-rich images guided by textual prompts. However, there are limitations to semantic understanding and commonsense reasoning in existing models when the input prompts are concise narrative, resulting in low-quality image generation. To improve the capacities for narrative prompts, we propose a simple-yet-effective parameter-efficient fine-tuning approach called the Semantic Understanding and Reasoning adapter (SUR-adapter) for pre-trained diffusion models. To reach this goal, we first collect and annotate a new dataset SURD which consists of more than 57,000 semantically corrected multi-modal samples. Each sample contains a simple narrative prompt, a complex keyword-based prompt, and a high-quality image. Then, we align the semantic representation of narrative prompts to the complex prompts and transfer knowledge of large language models (LLMs) to our SUR-adapter via knowledge distillation so that it can acquire the powerful semantic understanding and reasoning capabilities to build a high-quality textual semantic representation for text-to-image generation. We conduct experiments by integrating multiple LLMs and popular pre-trained diffusion models to show the effectiveness of our approach in enabling diffusion models to understand and reason concise natural language without image quality degradation. Our approach can make text-to-image diffusion models easier to use with better user experience, which demonstrates our approach has the potential for further advancing the development of user-friendly text-to-image generation models by bridging the semantic gap between simple narrative prompts and complex keyword-based prompts. The code is released at https://github.com/Qrange-group/SUR-adapter.

In diffusion models, UNet is the most popular network backbone, since its long skip connects (LSCs) to connect distant network blocks can aggregate long-distant information and alleviate vanishing gradient. Unfortunately, UNet often suffers from unstable training in diffusion models which can be alleviated by scaling its LSC coefficients smaller. However, theoretical understandings of the instability of UNet in diffusion models and also the performance improvement of LSC scaling remain absent yet. To solve this issue, we theoretically show that the coefficients of LSCs in UNet have big effects on the stableness of the forward and backward propagation and robustness of UNet. Specifically, the hidden feature and gradient of UNet at any layer can oscillate and their oscillation ranges are actually large which explains the instability of UNet training. Moreover, UNet is also provably sensitive to perturbed input, and predicts an output distant from the desired output, yielding oscillatory loss and thus oscillatory gradient. Besides, we also observe the theoretical benefits of the LSC coefficient scaling of UNet in the stableness of hidden features and gradient and also robustness. Finally, inspired by our theory, we propose an effective coefficient scaling framework ScaleLong that scales the coefficients of LSC in UNet and better improve the training stability of UNet. Experimental results on CIFAR10, CelebA, ImageNet and COCO show that our methods are superior to stabilize training, and yield about 1.5x training acceleration on different diffusion models with UNet or UViT backbones.

Neural fields have achieved impressive advancements in view synthesis and scene reconstruction. However, editing these neural fields remains challenging due to the implicit encoding of geometry and texture information. In this paper, we propose DreamEditor, a novel framework that enables users to perform controlled editing of neural fields using text prompts. By representing scenes as mesh-based neural fields, DreamEditor allows localized editing within specific regions. DreamEditor utilizes the text encoder of a pretrained text-to-Image diffusion model to automatically identify the regions to be edited based on the semantics of the text prompts. Subsequently, DreamEditor optimizes the editing region and aligns its geometry and texture with the text prompts through score distillation sampling [Poole et al. 2022]. Extensive experiments have demonstrated that DreamEditor can accurately edit neural fields of real-world scenes according to the given text prompts while ensuring consistency in irrelevant areas. DreamEditor generates highly realistic textures and geometry, significantly surpassing previous works in both quantitative and qualitative evaluations.

Deep learning-based hyperspectral image (HSI) super-resolution, which aims to generate high spatial resolution HSI (HR-HSI) by fusing hyperspectral image (HSI) and multispectral image (MSI) with deep neural networks (DNNs), has attracted lots of attention. However, neural networks require large amounts of training data, hindering their application in real-world scenarios. In this letter, we propose a novel adversarial automatic data augmentation framework ADASR that automatically optimizes and augments HSI-MSI sample pairs to enrich data diversity for HSI-MSI fusion. Our framework is sample-aware and optimizes an augmentor network and two downsampling networks jointly by adversarial learning so that we can learn more robust downsampling networks for training the upsampling network. Extensive experiments on two public classical hyperspectral datasets demonstrate the effectiveness of our ADASR compared to the state-of-the-art methods.

Matching clothing images from customers and online shopping stores has rich applications in e-commerce. Existing algorithms mostly encode an image as a global feature vector and perform retrieval via global representation matching. However, distinctive local information on clothing is immersed in this global representation, resulting in sub-optimized performance. To address this issue, we propose a novel graph reasoning network (GRNet) on a similarity pyramid, which learns similarities between a query and a gallery cloth by using both initial pairwise multi-scale feature representations and matching propagation for unaligned representations. The query local representations at each scale are aligned with those of the gallery via an adaptive window pooling module. The similarity pyramid is represented by a similarity graph, where nodes represent similarities between clothing components at different scales, and the final matching score is obtained by message propagation along edges. In GRNet, graph reasoning is solved by training a graph convolutional network, enabling the alignment of salient clothing components to improve clothing retrieval. To facilitate future research, we introduce a new benchmark, i.e. FindFashion, containing rich annotations of bounding boxes, views, occlusions, and cropping. Extensive experiments show that GRNet obtains new state-of-the-art results on three challenging benchmarks, e.g. pushing the accuracy of top-1, top-20, and top-50 on DeepFashion to 27, 66, and 75 percent (i.e. 6, 12, and 10 percent absolute improvements), outperforming competitors with large margins. On FindFashion, GRNet achieves considerable improvements on all empirical settings.

Textual logical reasoning, especially question-answering (QA) tasks with logical reasoning, requires awareness of particular logical structures. The passage-level logical relations represent entailment or contradiction between propositional units (e.g., a concluding sentence). However, such structures are unexplored as current QA systems focus on entity-based relations. In this work, we propose logic structural-constraint modeling to solve the logical reasoning QA and introduce discourse-aware graph networks (DAGNs). The networks first construct logic graphs leveraging in-line discourse connectives and generic logic theories, then learn logic representations by end-to-end evolving the logic relations with an edge-reasoning mechanism and updating the graph features. This pipeline is applied to a general encoder, whose fundamental features are joined with the high-level logic features for answer prediction. Experiments on three textual logical reasoning datasets demonstrate the reasonability of the logical structures built in DAGNs and the effectiveness of the learned logic features. Moreover, zero-shot transfer results show the features' generality to unseen logical texts.

Controlling a non-statically bipedal robot is challenging due to the complex dynamics and multi-criterion optimization involved. Recent works have demonstrated the effectiveness of deep reinforcement learning (DRL) for simulation and physical robots. In these methods, the rewards from different criteria are normally summed to learn a scalar function. However, a scalar is less informative and may be insufficient to derive effective information for each reward channel from the complex hybrid rewards. In this work, we propose a novel reward-adaptive reinforcement learning method for biped locomotion, allowing the control policy to be simultaneously optimized by multiple criteria using a dynamic mechanism. The proposed method applies a multi-head critic to learn a separate value function for each reward component, leading to hybrid policy gradients. We further propose dynamic weight, allowing each component to optimize the policy with different priorities. This hybrid and dynamic policy gradient (HDPG) design makes the agent learn more efficiently. We show that the proposed method outperforms summed-up-reward approaches and is able to transfer to physical robots. The MuJoCo results further demonstrate the effectiveness and generalization of HDPG.

Vision-and-language navigation (VLN) asks an agent to follow a given language instruction to navigate through a real 3D environment. Despite significant advances, conventional VLN agents are trained typically under disturbance-free environments and may easily fail in real-world navigation scenarios, since they are unaware of how to deal with various possible disturbances, such as sudden obstacles or human interruptions, which widely exist and may usually cause an unexpected route deviation. In this paper, we present a model-agnostic training paradigm, called Progressive Perturbation-aware Contrastive Learning (PROPER) to enhance the generalization ability of existing VLN agents to the real world, by requiring them to learn towards deviation-robust navigation. Specifically, a simple yet effective path perturbation scheme is introduced to implement the route deviation, with which the agent is required to still navigate successfully following the original instruction. Since directly enforcing the agent to learn perturbed trajectories may lead to insufficient and inefficient training, a progressively perturbed trajectory augmentation strategy is designed, where the agent can self-adaptively learn to navigate under perturbation with the improvement of its navigation performance for each specific trajectory. For encouraging the agent to well capture the difference brought by perturbation and adapt to both perturbation-free and perturbation-based environments, a perturbation-aware contrastive learning mechanism is further developed by contrasting perturbation-free trajectory encodings and perturbation-based counterparts. Extensive experiments on the standard Room-to-Room (R2R) benchmark show that PROPER can benefit multiple state-of-the-art VLN baselines in perturbation-free scenarios. We further collect the perturbed path data to construct an introspection subset based on the R2R, called Path-Perturbed R2R (PP-R2R). The results on PP-R2R show unsatisfying robustness of popular VLN agents and the capability of PROPER in improving the navigation robustness under deviation.

Medical diagnosis assistant (MDA) aims to build an interactive diagnostic agent to sequentially inquire about symptoms for discriminating diseases. However, since the dialogue records for building a patient simulator are collected passively, the collected records might be deteriorated by some task-unrelated biases, such as the preference of the collectors. These biases might hinder the diagnostic agent to capture transportable knowledge from the simulator. This work identifies and resolves two representative non-causal biases, i.e., (i) default-answer bias and (ii) distributional inquiry bias. Specifically, Bias (i) originates from the patient simulator which tries to answer the unrecorded inquiries with some biased default answers. To eliminate this bias and improve upon a well-known causal inference technique, i.e., propensity score matching, we propose a novel propensity latent matching in building a patient simulator to effectively answer unrecorded inquiries; Bias (ii) inherently comes along with the passively collected data that the agent might learn by remembering what to inquire within the training data while not able to generalize to the out-of-distribution cases. To this end, we propose a progressive assurance agent, which includes the dual processes accounting for symptom inquiry and disease diagnosis respectively. The diagnosis process pictures the patient mentally and probabilistically by intervention to eliminate the effect of the inquiry behavior. And the inquiry process is driven by the diagnosis process to inquire about symptoms to enhance the diagnostic confidence which alters as the patient distribution changes. In this cooperative manner, our proposed agent can improve upon the out-of-distribution generalization significantly. Extensive experiments demonstrate that our framework achieves new state-of-the-art performance and possesses the advantage of transportability.

Existing visual question answering methods often suffer from cross-modal spurious correlations and oversimplified event-level reasoning processes that fail to capture event temporality, causality, and dynamics spanning over the video. In this work, to address the task of event-level visual question answering, we propose a framework for cross-modal causal relational reasoning. In particular, a set of causal intervention operations is introduced to discover the underlying causal structures across visual and linguistic modalities. Our framework, named Cross-Modal Causal RelatIonal Reasoning (CMCIR), involves three modules: i) Causality-aware Visual-Linguistic Reasoning (CVLR) module for collaboratively disentangling the visual and linguistic spurious correlations via front-door and back-door causal interventions; ii) Spatial-Temporal Transformer (STT) module for capturing the fine-grained interactions between visual and linguistic semantics; iii) Visual-Linguistic Feature Fusion (VLFF) module for learning the global semantic-aware visual-linguistic representations adaptively. Extensive experiments on four event-level datasets demonstrate the superiority of our CMCIR in discovering visual-linguistic causal structures and achieving robust event-level visual question answering.

Metro origin-destination prediction is a crucial yet challenging time-series analysis task in intelligent transportation systems, which aims to accurately forecast two specific types of cross-station ridership, i.e., Origin-Destination (OD) one and Destination-Origin (DO) one. However, complete OD matrices of previous time intervals can not be obtained immediately in online metro systems, and conventional methods only used limited information to forecast the future OD and DO ridership separately. In this work, we proposed a novel neural network module termed Heterogeneous Information Aggregation Machine (HIAM), which fully exploits heterogeneous information of historical data (e.g., incomplete OD matrices, unfinished order vectors, and DO matrices) to jointly learn the evolutionary patterns of OD and DO ridership. Specifically, an OD modeling branch estimates the potential destinations of unfinished orders explicitly to complement the information of incomplete OD matrices, while a DO modeling branch takes DO matrices as input to capture the spatial-temporal distribution of DO ridership. Moreover, a Dual Information Transformer is introduced to propagate the mutual information among OD features and DO features for modeling the OD-DO causality and correlation. Based on the proposed HIAM, we develop a unified Seq2Seq network to forecast the future OD and DO ridership simultaneously. Extensive experiments conducted on two large-scale benchmarks demonstrate the effectiveness of our method for online metro origin-destination prediction. Our code is resealed at https://github.com/HCPLab-SYSU/HIAM.

Aiming at recognizing and localizing objects of novel categories with just a few reference samples, few-shot object detection (FSOD) is quite a challenging task. Previous works rely heavily on the fine-tuning process to transfer their models to the novel categories. They are flawed in the real application since the fine-tuning process is time-consuming and it suffers from serious deterioration on the low-quality support set. Based on the observation, this paper proposes an instant-response and accurate few-shot object detector (IRA-FSOD) that can detect the objects from novel categories without fine-tuning. We carefully analyze the limitations of widely-used Faster R-CNN and transform it to IRA-FSOD. Specifically, we first propose a novel semi-supervised Region Proposal Network (SS-RPN) module and a switch classifier module to precisely recognize the potential foreground instances from novel categories without fine-tuning. Moreover, we introduce two explicit inference strategies into the localization module, including explicit localization score and semi-explicit box regression, to alleviate over-fitting towards the base categories. Extensive experiments demonstrates that the proposed IRA-FSOD not only accomplish few-shot object detection with the instant-response, but also reaches state-of-the-art performance under various FSOD protocols and settings.

Electricity theft is the primary cause of electrical losses in power systems, which severely harms the economic benefits of electricity providers and threatens the safety of the power supply. However, due to the inherent complex correlation and periodicity of electricity consumption and the low efficiency of large-scale data processing, detecting anomalies in electricity consumption data accurately and efficiently remains challenging. Existing methods usually focus on first-order information and ignore the second-order representation learning that can efficiently model global temporal dependency and facilitate discriminative representation learning of electricity consumption data. In this article, we propose a novel electricity theft detection framework named hybrid-order representation learning network (HORLN). Specifically, the sequential electricity consumption data is transformed into the matrix format containing weekly consumption records. Then, an inter-and-intra week convolution block is designed to capture multiscale features in a local-to-global manner. Meanwhile, a self-dependency modeling module is proposed to learn the second-order representations from self-correlation matrices, which are finally combined with the first-order representations to predict the anomaly scores of electricity consumers. Extensive experiments on a real-world benchmark demonstrate the advantages of our HORLN over state-of-the-art methods.

Due to the difficulty of collecting paired Low-Resolution (LR) and High-Resolution (HR) images, the recent research on single image Super-Resolution (SR) has often been criticized for the data bottleneck of the synthetic image degradation between LRs and HRs. Recently, the emergence of real-world SR datasets, e.g., RealSR and DRealSR, promotes the exploration of Real-World image Super-Resolution (RWSR). RWSR exposes a more practical image degradation, which greatly challenges the learning capacity of deep neural networks to reconstruct high-quality images from low-quality images collected in realistic scenarios. In this paper, we explore Taylor series approximation in prevalent deep neural networks for image reconstruction, and propose a very general Taylor architecture to derive Taylor Neural Networks (TNNs) in a principled manner. Our TNN builds Taylor Modules with Taylor Skip Connections (TSCs) to approximate the feature projection functions, following the spirit of Taylor Series. TSCs introduce the input connected directly with each layer at different layers, to sequentially produces different high-order Taylor maps to attend more image details, and then aggregate the different high-order information from different layers. Only via simple skip connections, TNN is compatible with various existing neural networks to effectively learn high-order components of the input image with little increase of parameters. Furthermore, we have conducted extensive experiments to evaluate our TNNs in different backbones on two RWSR benchmarks, which achieve a superior performance in comparison with existing baseline methods.

Factorization machines (FMs) and their neural network variants (neural FMs) for modeling second-order feature interactions are effective in building modern recommendation systems. However, feature interactions are based upon pairs of features, whereas multi-features correlations commonly arise in real-world financial product recommendation scenarios. We propose an effective neural recommender system, graph-convolved factorization machine (GCFM), with the spirit of the symbolic graph reasoning principle that provides lightweight and interpretable recommendation suggestions. Given a sample for the recommendation, GCFM constructs the corresponding dense feature embeddings and computes the sample-specific feature relationship graph. Then, a multi-filter graph-convolved feature crossing (GCFC) layer for feature embeddings establishes cross features with their neighboring embeddings. GCFM thus extends the feature interactions from pairs to neighbors to capture more comprehensive and explainable information while simultaneously reaping the advantages of representation learning. To exploit these capabilities, we apply a Graph Bayesian Optimization (GBO). During training, our GBO automatically optimizes our GCFM, including training hyperparameters and architecture hyperparameters. Besides, we conduct extensive experiments on two public financial applications benchmarks, USCFC and OTC, and two real-world datasets that we collect offline. Our GCFM significantly outperforms state-of-the-art algorithms and shows its interpretability in recommendation tasks.

Real-world image super-resolution is a practical image restoration problem that aims to obtain high-quality images from in-the-wild input, has recently received considerable attention with regard to its tremendous application potentials. Although deep learning-based methods have achieved promising restoration quality on real-world image super-resolution datasets, they ignore the relationship between L1- and perceptual- minimization and roughly adopt auxiliary large-scale datasets for pre-training. In this paper, we discuss the image types within a corrupted image and the property of perceptual- and Euclidean- based evaluation protocols. Then we propose a method, Real-World image Super-Resolution by Exclusionary Dual-Learning (RWSR-EDL) to address the feature diversity in perceptual- and L1- based cooperative learning. Moreover, a noise-guidance data collection strategy is developed to address the training time consumption in multiple datasets optimization. When an auxiliary dataset is incorporated, RWSR-EDL achieves promising results and repulses any training time increment by adopting the noise-guidance data collection strategy. Extensive experiments show that RWSR-EDL achieves competitive performance over state-of-the-art methods on four in-the-wild image super-resolution datasets.

Fine-grained face editing, as a special case of image translation task, aims at modifying face attributes according to users' preference. Although generative adversarial networks (GANs) have achieved great success in general image translation tasks, these models cannot be directly applied in the face editing problem. Ideal face editing is challenging as it has two special requirements -- personalization and spatial-awareness. To address these issues, we propose a novel Personalized Spatial-aware Affine Modulation (PSAM) method based on a general GAN structure. The key idea is to modulate the intermediate features in a personalized and spatial-aware manner, which corresponds to the face editing procedure. Specifically, for personalization, we adopt both the face image and the desired attribute as input to generate the modulation tensors. For spatial-aware, we set these tensors to be of the same size as the input image, allowing pixel-wise modulation. Extensive experiments in four fine-grained face editing tasks, i.e., makeup, expression, illumination and aging, demonstrate the effectiveness of the proposed PSAM method. The synthesis results of PSAM can be further boosted by a new transferable training strategy. To facilitate research of face editing, we also construct a new large-scale makeup dataset. The code and dataset are available at http://colalab.org/projects/PSAM.

In recent years, the Transformer architecture has shown its superiority in the video-based person re-identification task. Inspired by video representation learning, these methods mainly focus on designing modules to extract informative spatial and temporal features. However, they are still limited in extracting local attributes and global identity information, which are critical for the person re-identification task. In this paper, we propose a novel Multi-Stage Spatial-Temporal Aggregation Transformer (MSTAT) with two novel designed proxy embedding modules to address the above issue. Specifically, MSTAT consists of three stages to encode the attribute-associated, the identity-associated, and the attribute-identity-associated information from the video clips, respectively, achieving the holistic perception of the input person. We combine the outputs of all the stages for the final identification. In practice, to save the computational cost, the Spatial-Temporal Aggregation (STA) modules are first adopted in each stage to conduct the self-attention operations along the spatial and temporal dimensions separately. We further introduce the Attribute-Aware and Identity-Aware Proxy embedding modules (AAP and IAP) to extract the informative and discriminative feature representations at different stages. All of them are realized by employing newly designed self-attention operations with specific meanings. Moreover, temporal patch shuffling is also introduced to further improve the robustness of the model. Extensive experimental results demonstrate the effectiveness of the proposed modules in extracting the informative and discriminative information from the videos, and illustrate the MSTAT can achieve state-of-the-art accuracies on various standard benchmarks.

Image Virtual try-on aims at replacing the cloth on a personal image with a garment image (in-shop clothes), which has attracted increasing attention from the multimedia and computer vision communities. Prior methods successfully preserve the character of clothing images, however, occlusion remains a pernicious effect for realistic virtual try-on. In this work, we first present a comprehensive analysis of the occlusions and categorize them into two aspects: i) Inherent-Occlusion: the ghost of the former cloth still exists in the try-on image; ii) Acquired-Occlusion: the target cloth warps to the unreasonable body part. Based on the in-depth analysis, we find that the occlusions can be simulated by a novel semantically-guided mixup module, which can generate semantic-specific occluded images that work together with the try-on images to facilitate training a de-occlusion try-on (DOC-VTON) framework. Specifically, DOC-VTON first conducts a sharpened semantic parsing on the try-on person. Aided by semantics guidance and pose prior, various complexities of texture are selectively blending with human parts in a copy-and-paste manner. Then, the Generative Module (GM) is utilized to take charge of synthesizing the final try-on image and learning to de-occlusion jointly. In comparison to the state-of-the-art methods, DOC-VTON achieves better perceptual quality by reducing occlusion effects.

Land remote-sensing analysis is a crucial research in earth science. In this work, we focus on a challenging task of land analysis, i.e., automatic extraction of traffic roads from remote-sensing data, which has widespread applications in urban development and expansion estimation. Nevertheless, conventional methods either only utilized the limited information of aerial images, or simply fused multimodal information (e.g., vehicle trajectories), thus cannot well recognize unconstrained roads. To facilitate this problem, we introduce a novel neural network framework termed cross-modal message propagation network (CMMPNet), which fully benefits the complementary different modal data (i.e., aerial images and crowdsourced trajectories). Specifically, CMMPNet is composed of two deep autoencoders for modality-specific representation learning and a tailor-designed dual enhancement module for cross-modal representation refinement. In particular, the complementary information of each modality is comprehensively extracted and dynamically propagated to enhance the representation of another modality. Extensive experiments on three real-world benchmarks demonstrate the effectiveness of our CMMPNet for robust road extraction benefiting from blending different modal data, either using image and trajectory data or image and light detection and ranging (LiDAR) data. From the experimental results, we observe that the proposed approach outperforms current state-of-the-art methods by large margins. Our source code is resealed on the project page http://lingboliu.com/multimodal_road_extraction.html.

Training the multi-label image recognition models with partial labels, in which merely some labels are known while others are unknown for each image, is a considerably challenging and practical task. To address this task, current algorithms mainly depend on pre-training classification or similarity models to generate pseudo labels for the unknown labels. However, these algorithms depend on sufficient multi-label annotations to train the models, leading to poor performance especially with low known label proportion. In this work, we propose to blend category-specific representation across different images to transfer information of known labels to complement unknown labels, which can get rid of pre-training models and thus does not depend on sufficient annotations. To this end, we design a unified semantic-aware representation blending (SARB) framework that exploits instance-level and prototype-level semantic representation to complement unknown labels by two complementary modules: 1) an instance-level representation blending (ILRB) module blends the representations of the known labels in an image to the representations of the unknown labels in another image to complement these unknown labels. 2) a prototype-level representation blending (PLRB) module learns more stable representation prototypes for each category and blends the representation of unknown labels with the prototypes of corresponding labels to complement these labels. Extensive experiments on the MS-COCO, Visual Genome, Pascal VOC 2007 datasets show that the proposed SARB framework obtains superior performance over current leading competitors on all known label proportion settings, i.e., with the mAP improvement of 4.6%, 4.6%, 2.2% on these three datasets when the known label proportion is 10%. Codes are available at https://github.com/HCPLab-SYSU/HCP-MLR-PL.

Multi-label image recognition is a fundamental yet practical task because real-world images inherently possess multiple semantic labels. However, it is difficult to collect large-scale multi-label annotations due to the complexity of both the input images and output label spaces. To reduce the annotation cost, we propose a structured semantic transfer (SST) framework that enables training multi-label recognition models with partial labels, i.e., merely some labels are known while other labels are missing (also called unknown labels) per image. The framework consists of two complementary transfer modules that explore within-image and cross-image semantic correlations to transfer knowledge of known labels to generate pseudo labels for unknown labels. Specifically, an intra-image semantic transfer module learns image-specific label co-occurrence matrix and maps the known labels to complement unknown labels based on this matrix. Meanwhile, a cross-image transfer module learns category-specific feature similarities and helps complement unknown labels with high similarities. Finally, both known and generated labels are used to train the multi-label recognition models. Extensive experiments on the Microsoft COCO, Visual Genome and Pascal VOC datasets show that the proposed SST framework obtains superior performance over current state-of-the-art algorithms. Codes are available at https://github.com/HCPLab-SYSU/HCP-MLR-PL.

Recent advances in deep learning significantly boost the performance of salient object detection (SOD) at the expense of labeling larger-scale per-pixel annotations. To relieve the burden of labor-intensive labeling, deep unsupervised SOD methods have been proposed to exploit noisy labels generated by handcrafted saliency methods. However, it is still difficult to learn accurate saliency details from rough noisy labels. In this paper, we propose to learn saliency from synthetic but clean labels, which naturally has higher pixel-labeling quality without the effort of manual annotations. Specifically, we first construct a novel synthetic SOD dataset by a simple copy-paste strategy. Considering the large appearance differences between the synthetic and real-world scenarios, directly training with synthetic data will lead to performance degradation on real-world scenarios. To mitigate this problem, we propose a novel unsupervised domain adaptive SOD method to adapt between these two domains by uncertainty-aware self-training. Experimental results show that our proposed method outperforms the existing state-of-the-art deep unsupervised SOD methods on several benchmark datasets, and is even comparable to fully-supervised ones.

The resurgence of unsupervised learning can be attributed to the remarkable progress of self-supervised learning, which includes generative$(\mathcal{G})$and discriminative$(\mathcal{D})$models. In computer vision, the mainstream self-supervised learning algorithms are$\mathcal{D}$models. However, designing a$\mathcal{D}$model could be over-complicated; also, some studies hinted that a$\mathcal{D}$model might not be as general and interpretable as a$\mathcal{G}$model. In this paper, we switch from$\mathcal{D}$models to$\mathcal{G}$models using the classical auto-encoder$(AE)$. Note that a vanilla$\mathcal{G}$model was far less efficient than a$\mathcal{D}$model in self-supervised computer vision tasks, as it wastes model capability on overfitting semantic-agnostic high-frequency details. Inspired by perceptual learning that could use cross-view learning to perceive concepts and semantics11Following [26], we refer to semantics as visual concepts, e.g., a semantic-ware model indicates the model can perceive visual concepts, and the learned features are efficient in object recognition, detection, etc., we propose a novel$AE$that could learn semantic-aware representation via cross-view image reconstruction. We use one view of an image as the input and another view of the same image as the reconstruction target. This kind of$AE$has rarely been studied before, and the optimization is very difficult. To enhance learning ability and find a feasible solution, we propose a semantic aligner that uses geometric transformation knowledge to align the hidden code of$AE$to help optimization. These techniques significantly improve the representation learning ability of$AE$and make selfsupervised learning with$\mathcal{G}$models possible. Extensive experiments on many large-scale benchmarks (e.g., ImageNet, COCO 2017, and SYSU-30k) demonstrate the effectiveness of our methods. Code is available at https://github.com/wanggrun/Semantic-Aware-AE.

Due to the sophisticated imaging process, an identical scene captured by different cameras could exhibit distinct imaging patterns, introducing distinct proficiency among the super-resolution (SR) models trained on images from different devices. In this paper, we investigate a novel and practical task coded cross-device SR, which strives to adapt a real-world SR model trained on the paired images captured by one camera to low-resolution (LR) images captured by arbitrary target devices. The proposed task is highly challenging due to the absence of paired data from various imaging devices. To address this issue, we propose an unsupervised domain adaptation mechanism for real-world SR, named Dual ADversarial Adaptation (DADA), which only requires LR images in the target domain with available real paired data from a source camera. DADA employs the Domain-Invariant Attention (DIA) module to establish the basis of target model training even without HR supervision. Furthermore, the dual framework of DADA facilitates an Inter-domain Adversarial Adaptation (InterAA) in one branch for two LR input images from two domains, and an Intra-domain Adversarial Adaptation (IntraAA) in two branches for an LR input image. InterAA and IntraAA together improve the model transferability from the source domain to the target. We empirically conduct experiments under six$\text{Real} \rightarrow \text{Real}$adaptation settings among three different cameras, and achieve superior performance compared with existing state-of-the-art approaches. We also evaluate the proposed DADA to address the adaptation to the video camera, which presents a promising re-search topic to promote the wide applications of real-world super-resolution. Our source code is publicly available at https://github.com/lonelyhopeIDADA.

Continual learning is a challenging real-world problem for constructing a mature AI system when data are provided in a streaming fashion. Despite recent progress in continual classification, the researches of continual object detection are impeded by the diverse sizes and numbers of objects in each image. Different from previous works that tune the whole network for all tasks, in this work, we present a simple and flexible framework for continual object detection via pRotOtypical taSk corrElaTion guided gaTing mechAnism (ROSETTA). Concretely, a unified framework is shared by all tasks while task-aware gates are introduced to automatically select sub-models for specific tasks. In this way, various knowledge can be successively memorized by storing their corresponding sub-model weights in this system. To make ROSETTA automatically determine which experience is available and useful, a prototypical task correlation guided Gating Diversity Controller (GDC) is introduced to adaptively adjust the diversity of gates for the new task based on class-specific prototypes. GDC module computes class-to-class correlation matrix to depict the cross-task correlation, and hereby activates more exclusive gates for the new task if a significant domain gap is observed. Comprehensive experiments on COCO-VOC, KITTI-Kitchen, class-incremental detection on VOC and sequential learning of four tasks show that ROSETTA yields state-of-the-art performance on both task-based and class-based continual object detection.11Codes are available at: https://github.com/dkxocl/ROSSETA.

Geometry problem solving is a well-recognized testbed for evaluating the high-level multimodal reasoning capability of deep models.In most existing works, two main geometry problems: calculation and proving, are usually treated as two specific tasks, hindering a deep model to unify its reasoning capability on multiple math tasks.However, in essence, these two tasks have similar problem representations and overlapped math knowledge which can improve the understanding and reasoning ability of a deep model on both two tasks.Therefore, we construct a large-scale Unified Geometry problem benchmark, UniGeo, which contains 4,998 calculation problems and 9,543 proving problems.Each proving problem is annotated with a multi-step proof with reasons and mathematical expressions.The proof can be easily reformulated as a proving sequence that shares the same formats with the annotated program sequence for calculation problems.Naturally, we also present a unified multitask Geometric Transformer framework, Geoformer, to tackle calculation and proving problems simultaneously in the form of sequence generation, which finally shows the reasoning ability can be improved on both two tasks by unifying formulation.Furthermore, we propose a Mathematical Expression Pretraining (MEP) method that aims to predict the mathematical expressions in the problem solution, thus improving the Geoformer model.Experiments on the UniGeo demonstrate that our proposed Geoformer obtains state-of-the-art performance by outperforming task-specific model NGS with over 5.6% and 3.2% accuracies on calculation and proving problems, respectively.1

Aiming at recognizing the samples from novel categories with few reference samples, few-shot learning (FSL) is a challenging problem. We found that the existing works often build their few-shot model based on the image-level feature by mixing all local-level features, which leads to the discriminative location bias and information loss in local details. To tackle the problem, this paper returns the perspective to the local-level feature and proposes a series of local-level strategies. Specifically, we present (a) a local-agnostic training strategy to avoid the discriminative location bias between the base and novel categories, (b) a novel local-level similarity measure to capture the accurate comparison between local-level features, and (c) a local-level knowledge transfer that can synthesize different knowledge transfers from the base category according to different location features. Extensive experiments justify that our proposed local-level strategies can significantly boost the performance and achieve 2.8%–7.2% improvements over the baseline across different benchmark datasets, which also achieves the state-of-the-art accuracy.

Crowd counting is a fundamental yet challenging task that aims to automatically estimate the number of people in crowded scenes. Nowadays, with the rapid development of thermal and depth sensors, thermal images and depth maps become more accessible, which are proven to be beneficial information in boosting the performance of crowd counting. Consequently, we propose a Mutual Attention Transformer (MAT) module to fully leverage the complementary information of different modalities. Specifically, our MAT employs a cross-modal mutual attention mechanism to utilize the features of one modality to enhance the features of the other. Moreover, to improve performance by learning better visual representation and further exploiting modality-wise comple-mentarity, we design a self-supervised pre-training method based on cross-modal image reconstruction. Extensive experiments on two standard benchmarks (i.e., RGBT-CC and ShanghaiTechRGBD) show that the proposed method is effective and universal for multimodal crowd counting, outper-forming previous state-of-the-art methods.

Compared with the well-explored cross-domain image recognition, cross-domain action recognition is a more challenging task because not only spatial but also temporal domain gaps exist across domains. Previous works attempt to bridge the temporal domain gap by aligning the domain-related key segments of videos from source and target domains. However, such practice overlooks the heterogeneous temporal domain gaps among different categories and presents temporal alignment strategies in a class-irrelevant manner. To address this issue, we propose to achieve class-wise temporal alignment for cross-domain action recognition via prototypical graph alignment (PGA). Concretely, we generate segment-level prototypes for the classes of both domains to capture per-class temporal dynamics. Furthermore, intra-domain and inter-domain prototypical graphs are established to mine the temporal relationships between each input video and its corresponding intra-domain and inter-domain prototypes. In this way, a discriminative and domain adaptive video representation is obtained by holistically reasoning cross-domain temporal dynamics. To class-wisely align the cross-domain video representations, each action category is equipped with a customized class-specific domain discriminator for temporal alignment via adversarial learning. Extensive experiments on three benchmarks show that PGA yeilds state-of-the-art performance on the task of cross-domain action recognition.

In Open Set Domain Adaptation (OSDA), large amounts of target samples are drawn from the implicit categories that never appear in the source domain. Due to the lack of their specific belonging, existing methods indiscriminately regard them as a single class “unknown”. We challenge this broadly-adopted practice that may arouse unexpected detrimental ef-fects because the decision boundaries between the implicit categories have been fully ignored. Instead, we propose Self-supervised Class-Discovering Adapter (SCDA) that attempts to achieve OSDA by gradually discovering those implicit classes, then incorporating them to restructure the classifier and update the domain-adaptive features iteratively. SCDA performs two alternate steps to achieve implicit class discov-ery and self-supervised OSDA, respectively. By jointly op-timizing for two tasks, SCDA achieves the state-of-the-art in OSDA and shows a competitive performance to unearth the implicit target classes.

In the semi-supervised object detection task, due to the scarcity of labeled data and the diversity and complexity of objects to be detected, the quality of pseudo-labels generated by existing methods for unlabeled data is relatively low, which severely restricts the performance of semi-supervised object detection. In this paper, we revisit the pseudo-labeling based Teacher-Student mutual learning framework for semi-supervised object detection and identify that the inconsistency of the location and feature of the candidate object proposals between the Teacher and the Student branches are the fatal cause of the low quality of the pseudo labels. To address this issue, we propose a simple yet effective technique within the mainstream teacher-student framework, called Double Check Soft Teacher, to overcome the harm caused by insufficient quality of pseudo labels. Specifically, our proposed method leverages teacher model to generate pseudo labels for the student model. Especially, the candidate boxes generated by the student model based on the pseudo label will be sent to the teacher model for "double check", and then the teacher model will output probabilistic soft label with background class for those candidate boxes, which will be used to train the student model. Together with a pseudo labeling mechanism based on the sum of the TOP-K prediction score, which improves the recall rate of pseudo labels, Double Check Soft Teacher consistently surpasses state-of-the-art methods by significant margins on the MS-COCO benchmark, pushing the new state-of-the-art. Source codes are available at https://github.com/wkfdb/DCST.

3D Garment modeling is a critical and challenging topic in the area of computer vision and graphics, with increasing attention focused on garment representation learning, garment reconstruction, and controllable garment manipulation, whereas existing methods were constrained to model garments under specific categories or with relatively simple topologies. In this paper, we propose a novel Neural Sewing Machine (NSM), a learning-based framework for structure-preserving 3D garment modeling, which is capable of learning representations for garments with diverse shapes and topologies and is successfully applied to 3D garment reconstruction and controllable manipulation. To model generic garments, we first obtain sewing pattern embedding via a unified sewing pattern encoding module, as the sewing pattern can accurately describe the intrinsic structure and the topology of the 3D garment. Then we use a 3D garment decoder to decode the sewing pattern embedding into a 3D garment using the UV-position maps with masks. To preserve the intrinsic structure of the predicted 3D garment, we introduce an inner-panel structure-preserving loss, an inter-panel structure-preserving loss, and a surface-normal loss in the learning process of our framework. We evaluate NSM on the public 3D garment dataset with sewing patterns with diverse garment shapes and categories. Extensive experiments demonstrate that the proposed NSM is capable of representing 3D garments under diverse garment shapes and topologies, realistically reconstructing 3D garments from 2D images with the preserved structure, and accurately manipulating the 3D garment categories, shapes, and topologies, outperforming the state-of-the-art methods by a clear margin.

We investigate a practical domain adaptation task, called source-free domain adaptation (SFUDA), where the source pretrained model is adapted to the target domain without access to the source data. Existing techniques mainly leverage self-supervised pseudo-labeling to achieve class-wise global alignment [1] or rely on local structure extraction that encourages the feature consistency among neighborhoods [2]. While impressive progress has been made, both lines of methods have their own drawbacks – the “global” approach is sensitive to noisy labels while the “local” counterpart suffers from the source bias. In this paper, we present Divide and Contrast (DaC), a new paradigm for SFUDA that strives to connect the good ends of both worlds while bypassing their limitations. Based on the prediction confidence of the source model, DaC divides the target data into source-like and target-specific samples, where either group of samples is treated with tailored goals under an adaptive contrastive learning framework. Specifically, the source-like samples are utilized for learning global class clustering thanks to their relatively clean labels. The more noisy target-specific data are harnessed at the instance level for learning the intrinsic local structures. We further align the source-like domain with the target-specific samples using a memory bank-based Maximum Mean Discrepancy (MMD) loss to reduce the distribution mismatch. Extensive experiments on VisDA, Office-Home, and the more challenging DomainNet have verified the superior performance of DaC over current state-of-the-art approaches. The code is available at https://github.com/ZyeZhang/DaC.git.

Hyperspectral image (HSI) contains rich spectral information and spatial features, but the huge amount of data often leads to problems of low clustering accuracy and large computational complexity. In this letter, a new clustering method for HSI is proposed, which is named fast spectral embedded clustering based on structured graph learning (FSECSGL). First, the low-dimensional representation of data can be obtained to reduce the scale by the fast spectral embedded method. Then, we use the embedded data to learn an optimal similarity matrix by structured graph learning. Furthermore, the learning structure graph gives feedback to the original bipartite graph to generate better spectral embedded data. As a result, we can obtain a better similarity matrix and clustering result by iteration, which can overcome the limitation of$K$-means initialization. Experiments show that this method can obtain good clustering performance compared with other methods.

Recognizing multiple labels of an image is a practical yet challenging task, and remarkable progress has been achieved by searching for semantic regions and exploiting label dependencies. However, current works utilize RNN/LSTM to implicitly capture sequential region/label dependencies, which cannot fully explore mutual interactions among the semantic regions/labels and do not explicitly integrate label co-occurrences. In addition, these works require large amounts of training samples for each category, and they are unable to generalize to novel categories with limited samples. To address these issues, we propose a knowledge-guided graph routing (KGGR) framework, which unifies prior knowledge of statistical label correlations with deep neural networks. The framework exploits prior knowledge to guide adaptive information propagation among different categories to facilitate multi-label analysis and reduce the dependency of training samples. Specifically, it first builds a structured knowledge graph to correlate different labels based on statistical label co-occurrence. Then, it introduces the label semantics to guide learning semantic-specific features to initialize the graph, and it exploits a graph propagation network to explore graph node interactions, enabling learning contextualized image feature representations. Moreover, we initialize each graph node with the classifier weights for the corresponding label and apply another propagation network to transfer node messages through the graph. In this way, it can facilitate exploiting the information of correlated labels to help train better classifiers, especially for labels with limited training samples. We conduct extensive experiments on the traditional multi-label image recognition (MLR) and multi-label few-shot learning (ML-FSL) tasks and show that our KGGR framework outperforms the current state-of-the-art methods by sizable margins on the public benchmarks.

Facial expression recognition (FER) has received significant attention in the past decade with witnessed progress, but data inconsistencies among different FER datasets greatly hinder the generalization ability of the models learned on one dataset to another. Recently, a series of cross-domain FER algorithms (CD-FERs) have been extensively developed to address this issue. Although each declares to achieve superior performance, comprehensive and fair comparisons are lacking due to inconsistent choices of the source/target datasets and feature extractors. In this work, we first propose to construct a unified CD-FER evaluation benchmark, in which we re-implement the well-performing CD-FER and recently published general domain adaptation algorithms and ensure that all these algorithms adopt the same source/target datasets and feature extractors for fair CD-FER evaluations. Based on the analysis, we find that most of the current state-of-the-art algorithms use adversarial learning mechanisms that aim to learn holistic domain-invariant features to mitigate domain shifts. However, these algorithms ignore local features, which are more transferable across different datasets and carry more detailed content for fine-grained adaptation. Therefore, we develop a novel adversarial graph representation adaptation (AGRA) framework that integrates graph representation propagation with adversarial learning to realize effective cross-domain holistic-local feature co-adaptation. Specifically, our framework first builds two graphs to correlate holistic and local regions within each domain and across different domains, respectively. Then, it extracts holistic-local features from the input image and uses learnable per-class statistical distributions to initialize the corresponding graph nodes. Finally, two stacked graph convolution networks (GCNs) are adopted to propagate holistic-local features within each domain to explore their interaction and across different domains for holistic-local feature co-adaptation. In this way, the AGRA framework can adaptively learn fine-grained domain-invariant features and thus facilitate cross-domain expression recognition. We conduct extensive and fair comparisons on the unified evaluation benchmark and show that the proposed AGRA framework outperforms previous state-of-the-art methods.

Joint filtering mainly uses an additional guidance image as a prior and transfers its structures to the target image in the filtering process. Different from existing approaches that rely on local linear models or hand-designed objective functions to extract the structural information from the guidance image, we propose a new joint filtering method based on a spatially variant linear representation model (SVLRM), where the target image is linearly represented by the guidance image. However, learning SVLRMs for vision tasks is a highly ill-posed problem. To estimate the spatially variant linear representation coefficients, we develop an effective approach based on a deep convolutional neural network (CNN). As such, the proposed deep CNN (constrained by the SVLRM) is able to model the structural information of both the guidance and input images. We show that the proposed approach can be effectively applied to a variety of applications, including depth/RGB image upsampling and restoration, flash deblurring, natural image denoising, and scale-aware filtering. In addition, we show that the linear representation model can be extended to high-order representation models (e.g., quadratic and cubic polynomial representations). Extensive experimental results demonstrate that the proposed method performs favorably against the state-of-the-art methods that have been specifically designed for each task.

Prior highly-tuned image parsing models are usually studied in a certain domain with a specific set of semantic labels and can hardly be adapted into other scenarios (e.g.sharing discrepant label granularity) without extensive re-training. Learning a single universal parsing model by unifying label annotations from different domains or at various levels of granularity is a crucial but rarely addressed topic. This poses many fundamental learning challenges, e.g.discovering underlying semantic structures among different label granularity or mining label correlation across relevant tasks. To address these challenges, we propose a graph reasoning and transfer learning framework, named "Graphonomy," which incorporates human knowledge and label taxonomy into the intermediate graph representation learning beyond local convolutions. In particular, Graphonomy learns the global and structured semantic coherency in multiple domains via semantic-aware graph reasoning and transfer, enforcing the mutual benefits of the parsing across domains (e.g.different datasets or co-related tasks). The Graphonomy includes two iterated modules: Intra-Graph Reasoning and Inter-Graph Transfer modules. The former extracts the semantic graph in each domain to improve the feature representation learning by propagating information with the graph; the latter exploits the dependencies among the graphs from different domains for bidirectional knowledge transfer. We apply Graphonomy to two relevant but different image understanding research topics: human parsing and panoptic segmentation, and show Graphonomy can handle both of them well via a standard pipeline against current state-of-the-art approaches. Moreover, some extra benefit of our framework is demonstrated, e.g., generating the human parsing at various levels of granularity by unifying annotations across different datasets.

Language instruction plays an essential role in the natural language grounded navigation tasks. However, navigators trained with limited human-annotated instructions may have difficulties in accurately capturing key information from the complicated instruction at different timesteps, leading to poor navigation performance. In this paper, we exploit to train a more robust navigator which is capable of dynamically extracting crucial factors from the long instruction, by using an adversarial attacking paradigm. Specifically, we propose a Dynamic Reinforced Instruction Attacker (DR-Attacker), which learns to mislead the navigator to move to the wrong target by destroying the most instructive information in instructions at different timesteps. By formulating the perturbation generation as a Markov Decision Process, DR-Attacker is optimized by the reinforcement learning algorithm to generate perturbed instructions sequentially during the navigation, according to a learnable attack score. Then, the perturbed instructions, which serve as hard samples, are used for improving the robustness of the navigator with an effective adversarial training strategy and an auxiliary self-supervised reasoning task. Experimental results on both Vision-and-Language Navigation (VLN) and Navigation from Dialog History (NDH) tasks show the superiority of our proposed method over state-of-the-art methods. Moreover, the visualization analysis shows the effectiveness of the proposed DR-Attacker, which can successfully attack crucial information in the instructions at different timesteps. Code is available at https://github.com/expectorlin/DR-Attacker.

Although deep convolutional neural networks (CNNs) have demonstrated remarkable performance on multiple computer vision tasks, researches on adversarial learning have shown that deep models are vulnerable to adversarial examples, which are crafted by adding visually imperceptible perturbations to the input images. Most of the existing adversarial attack methods only create a single adversarial example for the input, which just gives a glimpse of the underlying data manifold of adversarial examples. An attractive solution is to explore the solution space of the adversarial examples and generate a diverse bunch of them, which could potentially improve the robustness of real-world systems and help prevent severe security threats and vulnerabilities. In this paper, we present an effective method, called Hamiltonian Monte Carlo with Accumulated Momentum (HMCAM), aiming to generate a sequence of adversarial examples. To improve the efficiency of HMC, we propose a new regime to automatically control the length of trajectories, which allows the algorithm to move with adaptive step sizes along the search direction at different positions. Moreover, we revisit the reason for high computational cost of adversarial training under the view of MCMC and design a new generative method called Contrastive Adversarial Training (CAT), which approaches equilibrium distribution of adversarial examples with only few iterations by building from small modifications of the standard Contrastive Divergence (CD) and achieve a trade-off between efficiency and accuracy. Both quantitative and qualitative analysis on several natural image datasets and practical systems have confirmed the superiority of the proposed algorithm.

Beyond generating long and topic-coherent paragraphs in traditional captioning tasks, the medical image report composition task poses more task-oriented challenges by requiring both the highly accurate medical term diagnosis and multiple heterogeneous forms of information, including impression and findings. Current methods often generate the most common sentences due to dataset bias for the individual case, regardless of whether the sentences properly capture key entities and relationships. Such limitations severely hinder their applicability and generalization capability in medical report composition, where the most critical sentences lie in the descriptions of abnormal diseases that are relatively rare. Moreover, some medical terms appearing in one report are often entangled with each other and co-occurred, for example, symptoms associated with a specific disease. To enforce the semantic consistency of medical terms to be incorporated into the final reports and encourage the sentence generation for rare abnormal descriptions, we propose a novel framework that unifies template retrieval and sentence generation to handle both common and rare abnormality while ensuring the semantic coherency among the detected medical terms. Specifically, our approach exploits hybrid-knowledge co-reasoning: 1) explicit relationships among all abnormal medical terms to induce the visual attention learning and topic representation encoding for better topic-oriented symptoms descriptions and 2) adaptive generation mode that changes between the template retrieval and sentence generation according to a contextual topic encoder. The experimental results on two medical report benchmarks demonstrate the superiority of the proposed framework in terms of both human and metrics evaluation.

Video self-supervised learning is a challenging task, which requires significant expressive power from the model to leverage rich spatial-temporal knowledge and generate effective supervisory signals from large amounts of unlabeled videos. However, existing methods fail to increase the temporal diversity of unlabeled videos and ignore elaborately modeling multi-scale temporal dependencies in an explicit way. To overcome these limitations, we take advantage of the multi-scale temporal dependencies within videos and propose a novel video self-supervised learning framework named Temporal Contrastive Graph Learning (TCGL), which jointly models the inter-snippet and intra-snippet temporal dependencies for temporal representation learning with a hybrid graph contrastive learning strategy. Specifically, a Spatial-Temporal Knowledge Discovering (STKD) module is first introduced to extract motion-enhanced spatial-temporal representations from videos based on the frequency domain analysis of discrete cosine transform. To explicitly model multi-scale temporal dependencies of unlabeled videos, our TCGL integrates the prior knowledge about the frame and snippet orders into graph structures, i.e., the intra-/inter-snippet Temporal Contrastive Graphs (TCG). Then, specific contrastive learning modules are designed to maximize the agreement between nodes in different graph views. To generate supervisory signals for unlabeled videos, we introduce an Adaptive Snippet Order Prediction (ASOP) module which leverages the relational knowledge among video snippets to learn the global context representation and recalibrate the channel-wise features adaptively. Experimental results demonstrate the superiority of our TCGL over the state-of-the-art methods on large-scale action recognition and video retrieval benchmarks. The code is publicly available at https://github.com/YangLiu9208/TCGL.

Due to the widespread applications in real-world scenarios, metro ridership prediction is a crucial but challenging task in intelligent transportation systems. However, conventional methods either ignore the topological information of metro systems or directly learn on physical topology, and cannot fully explore the patterns of ridership evolution. To address this problem, we model a metro system as graphs with various topologies and propose a unified Physical-Virtual Collaboration Graph Network (PVCGN), which can effectively learn the complex ridership patterns from the tailor-designed graphs. Specifically, a physical graph is directly built based on the realistic topology of the studied metro system, while a similarity graph and a correlation graph are built with virtual topologies under the guidance of the inter-station passenger flow similarity and correlation. These complementary graphs are incorporated into a Graph Convolution Gated Recurrent Unit (GC-GRU) for spatial-temporal representation learning. Further, a Fully-Connected Gated Recurrent Unit (FC-GRU) is also applied to capture the global evolution tendency. Finally, we develop a Seq2Seq model with GC-GRU and FC-GRU to forecast the future metro ridership sequentially. Extensive experiments on two large-scale benchmarks (e.g., Shanghai Metro and Hangzhou Metro) well demonstrate the superiority of our PVCGN for station-level metro ridership prediction. Moreover, we apply the proposed PVCGN to address the online origin-destination (OD) ridership prediction and the experiment results show the universality of our method. Our code and benchmarks are available athttps://github.com/HCPLab-SYSU/PVCGN.

Referring image segmentation aims at segmenting out the object or stuff referred to by a natural language expression. The challenge of this task lies in the requirement of understanding both vision and language. The linguistic structure of a referring expression can provide an intuitive and explainable layout for reasoning over visual and linguistic concepts. In this paper, we propose a structured attention network (SANet) to explore the multimodal reasoning over the dependency tree parsed from the referring expression. Specifically, SANet implements the multimodal reasoning using an attentional multimodal tree-structure recurrent module (AMTreeGRU) in a bottom-up manner. In addition, for spatial detail improvement, SANet further incorporates the semantics-guided low-level features into high-level ones using the proposed attentional skip connection module. Extensive experiments on four public benchmark datasets demonstrate the superiority of our proposed SANet with more explainable visualization examples.

Though beneficial for encouraging the visual question answering (VQA) models to discover the underlying knowledge by exploiting the input-output correlation beyond image and text contexts, the existing knowledge VQA data sets are mostly annotated in a crowdsource way, e.g., collecting questions and external reasons from different users via the Internet. In addition to the challenge of knowledge reasoning, how to deal with the annotator bias also remains unsolved, which often leads to superficial overfitted correlations between questions and answers. To address this issue, we propose a novel data set named knowledge-routed visual question reasoning for VQA model evaluation. Considering that a desirable VQA model should correctly perceive the image context, understand the question, and incorporate its learned knowledge, our proposed data set aims to cut off the shortcut learning exploited by the current deep embedding models and push the research boundary of the knowledge-based visual question reasoning. Specifically, we generate the question-answer pair based on both the visual genome scene graph and an external knowledge base with controlled programs to disentangle the knowledge from other biases. The programs can select one or two triplets from the scene graph or knowledge base to push multistep reasoning, avoid answer ambiguity, and balance the answer distribution. In contrast to the existing VQA data sets, we further imply the following two major constraints on the programs to incorporate knowledge reasoning. First, multiple knowledge triplets can be related to the question, but only one knowledge relates to the image object. This can enforce the VQA model to correctly perceive the image instead of guessing the knowledge based on the given question solely. Second, all questions are based on different knowledge, but the candidate answers are the same for both the training and test sets. We make the testing knowledge unused during training to evaluate whether a model can understand question words and handle unseen combinations. Extensive experiments with various baselines and state-of-the-art VQA models are conducted to demonstrate that there still exists a big gap between the model with and without groundtruth supporting triplets when given the embedded knowledge base. This reveals the weakness of the current deep embedding models on the knowledge reasoning problem.

Current state-of-the-art visual recognition systems usually rely on the following pipeline: 1) pretraining a neural network on a large-scale data set (e.g., ImageNet) and 2) finetuning the network weights on a smaller, task-specific data set. Such a pipeline assumes that the sole weight adaptation is able to transfer the network capability from one domain to another domain based on a strong assumption that a fixed architecture is appropriate for all domains. However, each domain with a distinct recognition target may need different levels/paths of feature hierarchy, where some neurons may become redundant, and some others are reactivated to form new network structures. In this work, we prove that dynamically adapting network architectures tailored for each domain task along with weight finetuning benefits in both efficiency and effectiveness, compared to the existing image recognition pipeline that only tunes the weights regardless of the architecture. Our method can be easily generalized to an unsupervised paradigm by replacing supernet training with self-supervised learning in the source domain tasks and performing linear evaluation in the downstream tasks. This further improves the search efficiency of our method. Moreover, we also provide principled and empirical analysis to explain why our approach works by investigating the ineffectiveness of existing neural architecture search. We find that preserving the joint distribution of the network architecture and weights is of importance. This analysis not only benefits image recognition but also provides insights for crafting neural networks. Experiments on five representative image recognition tasks, such as person re-identification, age estimation, gender recognition, image classification, and unsupervised domain adaptation, demonstrate the effectiveness of our method.

Visual commonsense knowledge has received growing attention in the reasoning of long-tailed visual relationships biased in terms of object and relation labels. Most current methods typically collect and utilize external knowledge for visual relationships by following the fixed reasoning path of {subject, object → predicate} to facilitate the recognition of infrequent relationships. However, the knowledge incorporation for such fixed multidependent path suffers from the data set biased and exponentially grown combinations of object and relation labels and ignores the semantic gap between commonsense knowledge and real scenes. To alleviate this, we propose configurable graph reasoning (CGR) to decompose the reasoning path of visual relationships and the incorporation of external knowledge, achieving configurable knowledge selection and personalized graph reasoning for each relation type in each image. Given a commonsense knowledge graph, CGR learns to match and retrieve knowledge for different subpaths and selectively compose the knowledge routed path. CGR adaptively configures the reasoning path based on the knowledge graph, bridges the semantic gap between the commonsense knowledge, and the real-world scenes and achieves better knowledge generalization. Extensive experiments show that CGR consistently outperforms previous state-of-the-art methods on several popular benchmarks and works well with different knowledge graphs. Detailed analyses demonstrated that CGR learned explainable and compelling configurations of reasoning paths.

Without prohibitive and laborious 3D annotations, weakly-supervised 3D human pose methods mainly employ the model regularization with geometric projection consistency or geometry estimation from multi-view images. Nevertheless, those approaches explicitly need known parameters of calibrated cameras, exhibiting a limited model generalization in various realistic scenarios. To mitigate this issue, in this paper, we propose a Deductive Weakly-Supervised Learning (DWSL) for 3D human pose machine. Our DWSL firstly learns latent representations on depth and camera pose for 3D pose reconstruction. Since weak supervision usually causes ill-conditioned learning or inferior estimation, our DWSL introduces deductive reasoning to make an inference for the human pose from a view to another and develops a reconstruction loss to demonstrate what the model learns and infers is reliable. This learning by deduction strategy employs the view-transform demonstration and structural rules derived from depth, geometry and angle constraints, which improves the reliability of the model training with weak supervision. On three 3D human pose benchmarks, we conduct extensive experiments to evaluate our proposed method, which achieves superior performance in comparison with state-of-the-art weak-supervised methods. Particularly, our model shows an appealing potential for learning from 2D data captured in dynamic outdoor scenes, which demonstrates promising robustness and generalization in realistic scenarios. Our code is publicly available at https://github.com/Xipeng-Chen/DWSL-3D-pose.

Human doctors with well-structured medical knowledge can diagnose a disease merely via a few conversations with patients about symptoms. In contrast, existing knowledge-grounded dialogue systems often require a large number of dialogue instances to learn as they fail to capture the correlations between different diseases and neglect the diagnostic experience shared among them. To address this issue, we propose a more natural and practical paradigm, i.e., low-resource medical dialogue generation, which can transfer the diagnostic experience from source diseases to target ones with a handful of data for adaptation. It is capitalized on a commonsense knowledge graph to characterize the prior disease-symptom relations. Besides, we develop a Graph-Evolving Meta-Learning (GEML) framework that learns to evolve the commonsense graph for reasoning disease-symptom correlations in a new disease, which effectively alleviates the needs of a large number of dialogues. More importantly, by dynamically evolving disease-symptom graphs, GEML also well addresses the real-world challenges that the disease-symptom correlations of each disease may vary or evolve along with more diagnostic cases. Extensive experiment results on the CMDD dataset and our newly-collected Chunyu dataset testify the superiority of our approach over state-of-the-art approaches. Besides, our GEML can generate an enriched dialogue-sensitive knowledge graph in an online manner, which could benefit other tasks grounded on knowledge graph.

Low-resource automatic speech recognition (ASR) is challenging, as the low-resource target language data cannot well train an ASR model. To solve this issue, meta-learning formulates ASR for each source language into many small ASR tasks and meta-learns a model initialization on all tasks from different source languages to access fast adaptation on unseen target languages. However, for different source languages, the quantity and difficulty vary greatly because of their different data scales and diverse phonological systems, which leads to task-quantity and task-difficulty imbalance issues and thus a failure of multilingual meta-learning ASR (MML-ASR). In this work, we solve this problem by developing a novel adversarial meta sampling (AMS) approach to improve MML-ASR. When sampling tasks in MML-ASR, AMS adaptively determines the task sampling probability for each source language. Specifically, for each source language, if the query loss is large, it means that its tasks are not well sampled to train ASR model in terms of its quantity and difficulty and thus should be sampled more frequently for extra learning. Inspired by this fact, we feed the historical task query loss of all source language domain into a network to learn a task sampling policy for adversarially increasing the current query loss of MML-ASR. Thus, the learnt task sampling policy can master the learning situation of each language and thus predicts good task sampling probability for each language for more effective learning. Finally, experiment results on two multilingual datasets show significant performance improvement when applying our AMS on MML-ASR, and also demonstrate the applicability of AMS to other low-resource speech tasks and transfer learning ASR approaches.

Jinghui Qin, Xiaodan Liang, Yining Hong, Jianheng Tang, Liang Lin. Proceedings of the 59th Annual Meeting of the Association for Computational Linguistics and the 11th International Joint Conference on Natural Language Processing (Volume 1: Long Papers). 2021.

Crowd counting is a fundamental yet challenging task, which desires rich information to generate pixel-wise crowd density maps. However, most previous methods only used the limited information of RGB images and cannot well discover potential pedestrians in unconstrained scenarios. In this work, we find that incorporating optical and thermal information can greatly help to recognize pedestrians. To promote future researches in this field, we introduce a large-scale RGBT Crowd Counting (RGBT-CC) benchmark, which contains 2,030 pairs of RGB-thermal images with 138,389 annotated people. Furthermore, to facilitate the multimodal crowd counting, we propose a cross-modal collaborative representation learning framework, which consists of multiple modality-specific branches, a modality-shared branch, and an Information Aggregation-Distribution Module (IADM) to capture the complementary information of different modalities fully. Specifically, our IADM incorporates two collaborative information transfers to dynamically enhance the modality-shared and modality-specific representations with a dual information propagation mechanism. Extensive experiments conducted on the RGBT-CC benchmark demonstrate the effectiveness of our framework for RGBT crowd counting. Moreover, the proposed approach is universal for multimodal crowd counting and is also capable to achieve superior performance on the ShanghaiTechRGBD [22] dataset. Finally, our source code and benchmark have been released at http://lingboliu.com/RGBT_Crowd_Counting.html.

Generalization on out-of-distribution (OOD) test data is an essential but underexplored topic in visual question answering. Current state-of-the-art VQA models often exploit the biased correlation between data and labels, which results in a large performance drop when the test and training data have different distributions. Inspired by the fact that humans can recognize novel concepts by composing existed concepts and capsule network’s ability of representing part-whole hierarchies, we propose to use capsules to represent parts and introduce "Linguistically Routing" to merge parts with human-prior hierarchies. Specifically, we first fuse visual features with a single question word as atomic parts. Then we introduce the "Linguistically Routing" to reweight the capsule connections between two layers such that: 1) the lower layer capsules can transfer their outputs to the most compatible higher capsules, and 2) two capsules can be merged if their corresponding words are merged in the question parse tree. The routing process maximizes the above unary and binary potentials across multiple layers and finally carves a tree structure inside the capsule network. We evaluate our proposed routing method on the CLEVR compositional generation test, the VQA-CP2 dataset and the VQAv2 dataset. The experimental results show that our proposed method can improve current VQA models on OOD split without losing performance on the in-domain test data.

Open-set semi-supervised learning (open-set SSL) investigates a challenging but practical scenario where out-of-distribution (OOD) samples are contained in the unlabeled data. While the mainstream technique seeks to completely filter out the OOD samples for semi-supervised learning (SSL), we propose a novel training mechanism that could effectively exploit the presence of OOD data for enhanced feature learning while avoiding its adverse impact on the SSL. We achieve this goal by first introducing a warm-up training that leverages all the unlabeled data, including both the in-distribution (ID) and OOD samples. Specifically, we perform a pretext task that enforces our feature extractor to obtain a high-level semantic understanding of the training images, leading to more discriminative features that can benefit the downstream tasks. Since the OOD samples are inevitably detrimental to SSL, we propose a novel cross-modal matching strategy to detect OOD samples. Instead of directly applying binary classification [39], we train the network to predict whether the data sample is matched to an assigned one-hot class label. The appeal of the proposed cross-modal matching over binary classification is the ability to generate a compatible feature space that aligns with the core classification task. Extensive experiments show that our approach substantially lifts the performance on open-set SSL and outperforms the state-of-the-art by a large margin.

Recently proposed neural architecture search (NAS) methods co-train billions of architectures in a supernet and estimate their potential accuracy using the network weights detached from the supernet. However, the ranking correlation between the architectures' predicted accuracy and their actual capability is incorrect, which causes the existing NAS methods' dilemma. We attribute this ranking correlation problem to the supernet training consistency shift, including feature shift and parameter shift. Feature shift is identified as dynamic input distributions of a hidden layer due to random path sampling. The input distribution dynamic affects the loss descent and finally affects architecture ranking. Parameter shift is identified as contradictory parameter updates for a shared layer lay in different paths in different training steps. The rapidly-changing parameter could not preserve architecture ranking. We address these two shifts simultaneously using a nontrivial supernet-∏ model, called ∏-NAS. Specifically, we employ a supernet-∏ model that contains cross-path learning to reduce the feature consistency shift between different paths. Meanwhile, we adopt a novel nontrivial mean teacher containing negative samples to overcome parameter shift and model collision. Furthermore, our ∏-NAS runs in an unsupervised manner, which can search for more transferable architectures. Extensive experiments on ImageNet and a wide range of downstream tasks (e.g., COCO 2017, ADE20K, and Cityscapes) demonstrate the effectiveness and universality of our ∏-NAS compared to supervised NAS. See Codes1.

Self-supervised learning (especially contrastive learning) has attracted great interest due to its huge potential in learning discriminative representations in an unsupervised manner. Despite the acknowledged successes, existing contrastive learning methods suffer from very low learning efficiency, e.g., taking about ten times more training epochs than supervised learning for comparable recognition accuracy. In this paper, we reveal two contradictory phenomena in contrastive learning that we call under-clustering and over-clustering problems, which are major obstacles to learning efficiency. Under-clustering means that the model cannot efficiently learn to discover the dissimilarity between inter-class samples when the negative sample pairs for contrastive learning are insufficient to differentiate all the actual object classes. Over-clustering implies that the model cannot efficiently learn features from excessive negative sample pairs, forcing the model to over-cluster samples of the same actual classes into different clusters. To simultaneously overcome these two problems, we propose a novel self-supervised learning framework using a truncated triplet loss. Precisely, we employ a triplet loss tending to maximize the relative distance between the positive pair and negative pairs to address the under-clustering problem; and we construct the negative pair by selecting a negative sample deputy from all negative samples to avoid the over-clustering problem, guaranteed by the Bernoulli Distribution model. We extensively evaluate our framework in several large-scale benchmarks (e.g., ImageNet, SYSU-30k, and COCO). The results demonstrate our model’s superiority (e.g., the learning efficiency) over the latest state-of-the-art methods by a clear margin. See Codes1.

The unsupervised representation learning for skeleton-based human action can be utilized in a variety of pose analysis applications. However, previous unsupervised methods focus on modeling the temporal dependencies in sequences, but take less effort in modeling the spatial structure in human action. To this end, we propose a novel unsupervised learning frame-work named Hierarchical Transformer for skeleton-based human action recognition. The Hierarchical Transformer consists of hierarchically aggregated self-attention modules for better capturing the spatial and temporal structure in the skeleton sequences. Furthermore, we propose to predict the motion between adjacent frames as a novel pre-training task for better capturing the long-term dependencies in sequences. Experimental results show that our method outperforms prior state-of-the-art unsupervised methods on NTU RGB+D and NW-UCLA datasets. Besides, our method also achieves state-of-the-art performance when the pre-trained model is transferred to SBU dataset, which demonstrates the generalizability of learned representation.

Network architecture search techniques have recently revolutionized massive application problems due to its advanced performance. However, previous works rely on a strong assumption that the optimizers are manually designed and fixed for all architectures, which deviates from the intuition that each architecture should be empowered with its optimal optimization strategy. In this paper, we propose to unify Dynamic Optimizer Search (DOS) and Network Architecture Search (NAS), aiming to jointly search for the optimization scheduling and network architecture, named as DOS-NAS. The scheduling process of DOS is formulated as a Markov Decision Process, optimized via the REINFORCE algorithm. Conjunct with NAS, our DOS-NAS learns a discrete temporal scheduling to dynamically configure optimizers from a predefined set, leading to personalized optimizer scheduling for the searched network. Extensive experiments on CIFAR-10, CIFAR-100 and Tiny-ImageNet demonstrate the validity of our proposed DOS-NAS in performance improvement and speeding up the convergence significantly. DOS-NAS provides a new paradigm for AutoML systems through the joint search over optimization strategy and network architecture.

Although deep reinforcement learning (RL) has been successfully applied to a variety of robotic control tasks, it’s still challenging to apply it to real-world tasks, due to the poor sample efficiency. Attempting to overcome this shortcoming, several works focus on reusing the collected trajectory data during the training by decomposing them into a set of policy-irrelevant discrete transitions. However, their improvements are somewhat marginal since i) the amount of the transitions is usually small, and ii) the value assignment only happens in the joint states. To address these issues, this paper introduces a concise yet powerful method to construct Continuous Transition, which exploits the trajectory information by exploiting the potential transitions along the trajectory. Specifically, we propose to synthesize new transitions for training by linearly interpolating the consecutive transitions. To keep the constructed transitions authentic, we also develop a discriminator to guide the construction process automatically. Extensive experiments demonstrate that our proposed method achieves a significant improvement in sample efficiency on various complex continuous robotic control problems in MuJoCo and outperforms the advanced model-based / model-free RL methods. The source code is available1.

Recognizing human emotion/expressions automatically is quite an expected ability for intelligent robotics, as it can promote better communication and cooperation with humans. Current deep-learning-based algorithms may achieve impressive performance in some lab-controlled environments, but they always fail to recognize the expressions accurately for the uncontrolled in-the-wild situation. Fortunately, facial action units (AU) describe subtle facial behaviors, and they can help distinguish uncertain and ambiguous expressions. In this work, we explore the correlations among the action units and facial expressions, and devise an AU-Expression Knowledge Constrained Representation Learning (AUE-CRL) framework to learn the AU representations without AU annotations and adaptively use representations to facilitate facial expression recognition. Specifically, it leverages AU-expression correlations to guide the learning of the AU classifiers, and thus it can obtain AU representations without incurring any AU annotations. Then, it introduces a knowledge-guided attention mechanism that mines useful AU representations under the constraint of AU-expression correlations. In this way, the framework can capture local discriminative and complementary features to enhance facial representation for facial expression recognition. We conduct experiments on the challenging uncontrolled datasets to demonstrate the superiority of the proposed framework over current state-of-the-art methods. Codes and trained models are available at https://github.com/HCPLab-SYSU/AUE-CRL.

In this paper, we introduce a novel task, referred to as Weakly-Supervised Spatio-Temporal Anomaly Detection (WSSTAD) in surveillance video. Specifically, given an untrimmed video, WSSTAD aims to localize a spatio-temporal tube (i.e., a sequence of bounding boxes at consecutive times) that encloses the abnormal event, with only coarse video-level annotations as supervision during training. To address this challenging task, we propose a dual-branch network which takes as input the proposals with multi-granularities in both spatial-temporal domains. Each branch employs a relationship reasoning module to capture the correlation between tubes/videolets, which can provide rich contextual information and complex entity relationships for the concept learning of abnormal behaviors. Mutually-guided Progressive Refinement framework is set up to employ dual-path mutual guidance in a recurrent manner, iteratively sharing auxiliary supervision information across branches. It impels the learned concepts of each branch to serve as a guide for its counterpart, which progressively refines the corresponding branch and the whole framework. Furthermore, we contribute two datasets, i.e., ST-UCF-Crime and STRA, consisting of videos containing spatio-temporal abnormal annotations to serve as the benchmarks for WSSTAD. We conduct extensive qualitative and quantitative evaluations to demonstrate the effectiveness of the proposed approach and analyze the key factors that contribute more to handle this task.

Recently deep neural networks (DNNs) have achieved significant success in real-world image super-resolution (SR). However, adversarial image samples with quasi-imperceptible noises could threaten deep learning SR models. In this paper, we propose a robust deep learning framework for real-world SR that randomly erases potential adversarial noises in the frequency domain of input images or features. The rationale is that on the SR task clean images or features have a different pattern from the attacked ones in the frequency domain. Observing that existing adversarial attacks usually add high-frequency noises to input images, we introduce a novel random frequency mask module that blocks out high-frequency components possibly containing the harmful perturbations in a stochastic manner. Since the frequency masking may not only destroys the adversarial perturbations but also affects the sharp details in a clean image, we further develop an adversarial sample classifier based on the frequency domain of images to determine if applying the proposed mask module. Based on the above ideas, we devise a novel real-world image SR framework that combines the proposed frequency mask modules and the proposed adversarial classifier with an existing super-resolution backbone network. Experiments show that our proposed method is more insensitive to adversarial attacks and presents more stable SR results than existing models and defenses.

In this paper, we focus on solving the navigation problem of embodied question answering (EmbodiedQA), where the lack of experience and common sense information essentially result in a failure finding target when the robot is spawn in unknown environments. We present a route planning method named Path Estimation and Memory Recalling (PEMR) framework. PEMR includes a “looking ahead” process, i.e. a visual feature extractor module that estimates feasible paths for gathering 3D navigational information; another process “looking behind” process that is a memory recalling mechanism aims at fully leveraging past experience collected by the feature extractor. To encourage the navigator to learn more accurate prior expert experience, we improve the original benchmark dataset and provide a family of evaluation metrics for diagnosing both navigation and question answering modules. We show strong experimental results of PEMR on the EmbodiedQA navigation task.

Channel pruning is a popular technique for compressing convolutional neural networks (CNNs), where various pruning criteria have been proposed to remove the redundant filters. From our comprehensive experiments, we found two blind spots of pruning criteria: (1) Similarity: There are some strong similarities among several primary pruning criteria that are widely cited and compared. According to these criteria, the ranks of filters’ Importance Score are almost identical, resulting in similar pruned structures. (2) Applicability: The filters' Importance Score measured by some pruning criteria are too close to distinguish the network redundancy well. In this paper, we analyze the above blind spots on different types of pruning criteria with layer-wise pruning or global pruning. We also break some stereotypes, such as that the results of $\ell_1$ and $\ell_2$ pruning are not always similar. These analyses are based on the empirical experiments and our assumption (Convolutional Weight Distribution Assumption) that the well-trained convolutional filters in each layer approximately follow a Gaussian-alike distribution. This assumption has been verified through systematic and extensive statistical tests.

Collaborative reasoning for understanding image-question pairs is a very critical but underexplored topic in interpretable visual question answering systems. Although very recent studies have attempted to use explicit compositional processes to assemble multiple subtasks embedded in questions, their models heavily rely on annotations or handcrafted rules to obtain valid reasoning processes, which leads to either heavy workloads or poor performance on compositional reasoning. In this paper, to better align image and language domains in diverse and unrestricted cases, we propose a novel neural network model that performs global reasoning on a dependency tree parsed from the question; thus, our model is called a parse-tree-guided reasoning network (PTGRN). This network consists of three collaborative modules: i) an attention module that exploits the local visual evidence of each word parsed from the question, ii) a gated residual composition module that composes the previously mined evidence, and iii) a parse-tree-guided propagation module that passes the mined evidence along the parse tree. Thus, PTGRN is capable of building an interpretable visual question answering (VQA) system that gradually derives image cues following question-driven parse-tree reasoning. Experiments on relational datasets demonstrate the superiority of PTGRN over current state-of-the-art VQA methods, and the visualization results highlight the explainable capability of our reasoning system.

Recently, deep convolutional neural networks have achieved significant success in salient object detection. However, existing state-of-the-art methods require high-end GPUs to achieve real-time performance, which makes it hard to adapt to low cost or portable devices. Although generic network architectures have been proposed to speed up inference on mobile devices, they are tailored to the task of image classification or semantic segmentation, and struggle to capture intrachannel and interchannel correlations that are essential for contrast modeling in salient object detection. Motivated by the above observations, we design a new deep-learning algorithm for fast salient object detection. The proposed algorithm for the first time achieves competitive accuracy and high inference efficiency simultaneously with a single CPU thread. Specifically, we propose a novel depthwise nonlocal module (DNL), which implicitly models contrast via harvesting intrachannel and interchannel correlations in a self-attention manner. In addition, we introduce a depthwise nonlocal network architecture that incorporates both DNLs module and inverted residual blocks. The experimental results show that our proposed network attains very competitive accuracy on a wide range of salient object detection datasets while achieving state-of-the-art efficiency among all existing deep-learning-based algorithms.

Human-object interaction detection that aims at detecting triplets is critical for the holistic human-centric scene understanding. Existing approaches ignore the modeling of correlations among hierarchical human parts and objects. In this work, we introduce a Hierarchical Reasoning Network (HRNet) to capture relations among human parts at multiple scales (including the holistic human, human region, and human keypoint levels) and objects via a unified graph. In particular, HRNet first constructs one multi-level human parts graph, each level of which consists of human parts at one specific scale, objects, and the unions of human part-object pairs as nodes, and their mutual visual and spatial layout relations as intra-level reasoning. To also capture the relations across scales, we further introduce inter-level reasoning between the nodes of two consecutive levels based on the prior of human body structure. The representations of graph nodes are propagated along intra-level and inter-level reasoning in turn during reasoning. Extensive experiments demonstrate our HRNet obtains new state-of-the-art results on three challenging HICO-DET, V-COCO and HOI-A benchmarks, validating the compelling effectiveness of the proposed method.

Existing vision-based action recognition is susceptible to occlusion and appearance variations, while wearable sensors can alleviate these challenges by capturing human motion with one-dimensional time-series signals (e.g. acceleration, gyroscope, and orientation). For the same action, the knowledge learned from vision sensors (videos or images) and wearable sensors, may be related and complementary. However, there exists a significantly large modality difference between action data captured by wearable-sensor and vision-sensor in data dimension, data distribution, and inherent information content. In this paper, we propose a novel framework, named Semantics-aware Adaptive Knowledge Distillation Networks (SAKDN), to enhance action recognition in vision-sensor modality (videos) by adaptively transferring and distilling the knowledge from multiple wearable sensors. The SAKDN uses multiple wearable-sensors as teacher modalities and uses RGB videos as student modalities. To preserve the local temporal relationship and facilitate employing visual deep learning models, we transform one-dimensional time-series signals of wearable sensors to two-dimensional images by designing a gramian angular field based virtual image generation model. Then, we introduce a novel Similarity-Preserving Adaptive Multi-modal Fusion Module (SPAMFM) to adaptively fuse intermediate representation knowledge from different teacher networks. Finally, to fully exploit and transfer the knowledge of multiple well-trained teacher networks to the student network, we propose a novel Graph-guided Semantically Discriminative Mapping (GSDM) module, which utilizes graph-guided ablation analysis to produce a good visual explanation to highlight the important regions across modalities and concurrently preserve the interrelations of original data. Experimental results on Berkeley-MHAD, UTD-MHAD, and MMAct datasets well demonstrate the effectiveness of our proposed SAKDN for adaptive knowledge transfer from wearable-sensors modalities to vision-sensors modalities. The code is publicly available at https://github.com/YangLiu9208/SAKDN.

Transferring human motion from a source to a target person poses great potential in computer vision and graphics applications. A crucial step is to manipulate sequential future motion while retaining the appearance characteristic. Previous work has either relied on crafted 3D human models or trained a separate model specifically for each target person, which is not scalable in practice. This work studies a more general setting, in which we aim to learn a single model to parsimoniously transfer motion from a source video to any target person given only one image of the person, named as Collaborative Parsing-Flow Network (CPF-Net). The paucity of information regarding the target person makes the task particularly challenging to faithfully preserve the appearance in varying designated poses. To address this issue, CPF-Net integrates the structured human parsing and appearance flow to guide the realistic foreground synthesis which is merged into the background by a spatio-temporal fusion module. In particular, CPF-Net decouples the problem into stages of human parsing sequence generation, foreground sequence generation and final video generation. The human parsing generation stage captures both the pose and the body structure of the target. The appearance flow is beneficial to keep details in synthesized frames. The integration of human parsing and appearance flow effectively guides the generation of video frames with realistic appearance. Finally, the dedicated designed fusion network ensure the temporal coherence. We further collect a large set of human dancing videos to push forward this research field. Both quantitative and qualitative results show our method substantially improves over previous approaches and is able to generate appealing and photo-realistic target videos given any input person image. All source code and dataset will be released at https://github.com/xiezhy6/CPF-Net.

Non-parallel text style transfer has attracted increasing research interests in recent years. Despite successes in transferring the style based on the encoder-decoder framework, current approaches still lack the ability to preserve the content and even logic of original sentences, mainly due to the large unconstrained model space or too simplified assumptions on latent embedding space. Since language itself is an intelligent product of humans with certain grammars and has a limited rule-based model space by its nature, relieving this problem requires reconciling the model capacity of deep neural networks with the intrinsic model constraints from human linguistic rules. To this end, we propose a method called Graph Transformer–based Auto-Encoder, which models a sentence as a linguistic graph and performs feature extraction and style transfer at the graph level, to maximally retain the content and the linguistic structure of original sentences. Quantitative experiment results on three non-parallel text style transfer tasks show that our model outperforms state-of-the-art methods in content preservation, while achieving comparable performance on transfer accuracy and sentence naturalness.

As a crucial component in intelligent transportation systems, traffic flow prediction has recently attracted widespread research interest in the field of artificial intelligence (AI) with the increasing availability of massive traffic mobility data. Its key challenge lies in how to integrate diverse factors (such as temporal rules and spatial dependencies) to infer the evolution trend of traffic flow. To address this problem, we propose a unified neural network called Attentive Traffic Flow Machine (ATFM), which can effectively learn the spatial-temporal feature representations of traffic flow with an attention mechanism. In particular, our ATFM is composed of two progressive Convolutional Long Short-Term Memory (ConvLSTM[1]) units connected with a convolutional layer. Specifically, the first ConvLSTM unit takes normal traffic flow features as input and generates a hidden state at each time-step, which is further fed into the connected convolutional layer for spatial attention map inference. The second ConvLSTM unit aims at learning the dynamic spatial-temporal representations from the attentionally weighted traffic flow features. Further, we develop two deep learning frameworks based on ATFM to predict citywide short-term/long-term traffic flow by adaptively incorporating the sequential and periodic data as well as other external influences. Extensive experiments on two standard benchmarks well demonstrate the superiority of the proposed method for traffic flow prediction. Moreover, to verify the generalization of our method, we also apply the customized framework to forecast the passenger pickup/dropoff demands in traffic prediction and show its superior performance. Our code and data are available athttps://github.com/liulingbo918/ATFM.

Multi-view clustering methods have been a focus in recent years because of their superiority in clustering performance. However, typical traditional multi-view clustering algorithms still have shortcomings in some aspects, such as removal of redundant information, utilization of various views and fusion of multi-view features. In view of these problems, this paper proposes a new multi-view clustering method, low-rank subspace multi-view clustering based on adaptive graph regularization. We construct two new data matrix decomposition models into a unified optimization model. In this framework, we address the significance of the common knowledge shared by the cross view and the unique knowledge of each view by presenting new low-rank and sparse constraints on the sparse subspace matrix. To ensure that we achieve effective sparse representation and clustering performance on the original data matrix, adaptive graph regularization and unsupervised clustering constraints are also incorporated in the proposed model to preserve the internal structural features of the data. Finally, the proposed method is compared with several state-of-the-art algorithms. Experimental results for five widely used multi-view benchmarks show that our proposed algorithm surpasses other state-of-the-art methods by a clear margin.

Significant progress has been made in recent years in image captioning, an active topic in the fields of vision and language. However, existing methods tend to yield overly general captions and consist of some of the most frequent words/phrases, resulting in inaccurate and indistinguishable descriptions (see Fig. 1). This is primarily due to (i) the conservative characteristic of traditional training objectives that drives the model to generate correct but hardly discriminative captions for similar images and (ii) the uneven word distribution of the ground-truth captions, which encourages generating highly frequent words/phrases while suppressing the less frequent but more concrete ones. In this work, we propose a novel global-local discriminative objective that is formulated on top of a reference model to facilitate generating fine-grained descriptive captions. Specifically, from a global perspective, we design a novel global discriminative constraint that pulls the generated sentence to better discern the corresponding image from all others in the entire dataset. From the local perspective, a local discriminative constraint is proposed to increase attention such that it emphasizes the less frequent but more concrete words/phrases, thus facilitating the generation of captions that better describe the visual details of the given images. We evaluate the proposed method on the widely used MS-COCO dataset, where it outperforms the baseline methods by a sizable margin and achieves competitive performance over existing leading approaches. We also conduct self-retrieval experiments to demonstrate the discriminability of the proposed method.

Existing deep reinforcement learning (RL) are devoted to research applications on video games, e.g., The Open Racing Car Simulator (TORCS) and Atari games. However, it remains under-explored for vision-based autonomous urban driving navigation (VB-AUDN). VB-AUDN requires a sophisticated agent working safely in structured, changing, and unpredictable environments; otherwise, inappropriate operations may lead to irreversible or catastrophic damages. In this work, we propose a deductive RL (DeRL) to address this challenge. A deduction reasoner (DR) is introduced to endow the agent with ability to foresee the future and to promote policy learning. Specifically, DR first predicts future transitions through a parameterized environment model. Then, DR conducts self-assessment at the predicted trajectory to perceive the consequences of current policy resulting in a more reliable decision-making process. Additionally, a semantic encoder module (SEM) is designed to extract compact driving representation from the raw images, which is robust to the changes of the environment. Extensive experimental results demonstrate that DeRL outperforms the state-of-the-art model-free RL approaches on the public CAR Learning to Act (CARLA) benchmark and presents a superior performance on success rate and driving safety for goal-directed navigation.

Person reidentification (Re-ID) benefits greatly from the accurate annotations of existing data sets (e.g., CUHK03 and Market-1501), which are quite expensive because each image in these data sets has to be assigned with a proper label. In this work, we ease the annotation of Re-ID by replacing the accurate annotation with inaccurate annotation, i.e., we group the images into bags in terms of time and assign a bag-level label for each bag. This greatly reduces the annotation effort and leads to the creation of a large-scale Re-ID benchmark called SYSU- 30k . The new benchmark contains 30k individuals, which is about 20 times larger than CUHK03 (1.3k individuals) and Market-1501 (1.5k individuals), and 30 times larger than ImageNet (1k categories). It sums up to 29606918 images. Learning a Re-ID model with bag-level annotation is called the weakly supervised Re-ID problem. To solve this problem, we introduce a differentiable graphical model to capture the dependencies from all images in a bag and generate a reliable pseudolabel for each person's image. The pseudolabel is further used to supervise the learning of the Re-ID model. Compared with the fully supervised Re-ID models, our method achieves state-of-the-art performance on SYSU- 30k and other data sets. The code, data set, and pretrained model will be available at https://github.com/wanggrun/SYSU-30k.

Few-shot learning aims to learn novel categories from very few samples given some base categories with sufficient training samples. The main challenge of this task is the novel categories are prone to dominated by color, texture, shape of the object or background context (namely specificity), which are distinct for the given few training samples but not common for the corresponding categories (see Figure 1). Fortunately, we find that transferring information of the correlated based categories can help learn the novel concepts and thus avoid the novel concept being dominated by the specificity. Besides, incorporating semantic correlations among different categories can effectively regularize this information transfer. In this work, we represent the semantic correlations in the form of structured knowledge graph and integrate this graph into deep neural networks to promote few-shot learning by a novel Knowledge Graph Transfer Network (KGTN). Specifically, by initializing each node with the classifier weight of the corresponding category, a propagation mechanism is learned to adaptively propagate node message through the graph to explore node interaction and transfer classifier information of the base categories to those of the novel ones. Extensive experiments on the ImageNet dataset show significant performance improvement compared with current leading competitors. Furthermore, we construct an ImageNet-6K dataset that covers larger scale categories, i.e, 6,000 categories, and experiments on this dataset further demonstrate the effectiveness of our proposed model.

Temporally language grounding in untrimmed videos is a newly-raised task in video understanding. Most of the existing methods suffer from inferior efficiency, lacking interpretability, and deviating from the human perception mechanism. Inspired by human's coarse-to-fine decision-making paradigm, we formulate a novel Tree-Structured Policy based Progressive Reinforcement Learning (TSP-PRL) framework to sequentially regulate the temporal boundary by an iterative refinement process. The semantic concepts are explicitly represented as the branches in the policy, which contributes to efficiently decomposing complex policies into an interpretable primitive action. Progressive reinforcement learning provides correct credit assignment via two task-oriented rewards that encourage mutual promotion within the tree-structured policy. We extensively evaluate TSP-PRL on the Charades-STA and ActivityNet datasets, and experimental results show that TSP-PRL achieves competitive performance over existing state-of-the-art methods.

We focus on Unsupervised Domain Adaptation (UDA) for the task of semantic segmentation. Recently, adversarial alignment has been widely adopted to match the marginal distribution of feature representations across two domains globally. However, this strategy fails in adapting the representations of the tail classes or small objects for semantic segmentation since the alignment objective is dominated by head categories or large objects. In contrast to adversarial alignment, we propose to explicitly train a domain-invariant classifier by generating and defensing against pointwise feature space adversarial perturbations. Specifically, we firstly perturb the intermediate feature maps with several attack objectives (i.e., discriminator and classifier) on each individual position for both domains, and then the classifier is trained to be invariant to the perturbations. By perturbing each position individually, our model treats each location evenly regardless of the category or object size and thus circumvents the aforementioned issue. Moreover, the domain gap in feature space is reduced by extrapolating source and target perturbed features towards each other with attack on the domain discriminator. Our approach achieves the state-of-the-art performance on two challenging domain adaptation tasks for semantic segmentation: GTA5 → Cityscapes and SYNTHIA → Cityscapes.

The success of DNNs has driven the extensive applications of person re-identification (ReID) into a new era. However, whether ReID inherits the vulnerability of DNNs remains unexplored. To examine the robustness of ReID systems is rather important because the insecurity of ReID systems may cause severe losses, e.g., the criminals may use the adversarial perturbations to cheat the CCTV systems. In this work, we examine the insecurity of current best-performing ReID models by proposing a learning-to-mis-rank formulation to perturb the ranking of the system output. As the cross-dataset transferability is crucial in the ReID domain, we also perform a back-box attack by developing a novel multi-stage network architecture that pyramids the features of different levels to extract general and transferable features for the adversarial perturbations. Our method can control the number of malicious pixels by using differentiable multi-shot sampling. To guarantee the inconspicuousness of the attack, we also propose a new perception loss to achieve better visual quality. Extensive experiments on four of the largest ReID benchmarks (i.e., Market1501, CUHK03, DukeMTMC, and MSMT17) not only show the effectiveness of our method, but also provides directions of the future improvement in the robustness of ReID systems. For example, the accuracy of one of the best-performing ReID systems drops sharply from 91.8% to 1.4% after being attacked by our method. Some attack results are shown in Fig. 1. The code is available at: https://github.com/whj363636/Adversarial-attack-on-Person-ReID-With-Deep-Mis-Ranking.

Neural Architecture Search (NAS), aiming at automatically designing network architectures by machines, is expected to bring about a new revolution in machine learning. Despite these high expectation, the effectiveness and efficiency of existing NAS solutions are unclear, with some recent works going so far as to suggest that many existing NAS solutions are no better than random architecture selection. The ineffectiveness of NAS solutions may be attributed to inaccurate architecture evaluation. Specifically, to speed up NAS, recent works have proposed under-training different candidate architectures in a large search space concurrently by using shared network parameters; however, this has resulted in incorrect architecture ratings and furthered the ineffectiveness of NAS. In this work, we propose to modularize the large search space of NAS into blocks to ensure that the potential candidate architectures are fully trained; this reduces the representation shift caused by the shared parameters and leads to the correct rating of the candidates. Thanks to the block-wise search, we can also evaluate all of the candidate architectures within each block. Moreover, we find that the knowledge of a network model lies not only in the network parameters but also in the network architecture. Therefore, we propose to distill the neural architecture (DNA) knowledge from a teacher model to supervise our block-wise architecture search, which significantly improves the effectiveness of NAS. Remarkably, the performance of our searched architectures has exceeded the teacher model, demonstrating the practicability of our method. Finally, our method achieves a state-of-the-art 78.4% top-1 accuracy on ImageNet in a mobile setting. All of our searched models along with the evaluation code are available at https://github.com/changlin31/DNA.

Recent researches on panoptic segmentation resort to a single end-to-end network to combine the tasks of instance segmentation and semantic segmentation. However, prior models only unified the two related tasks at the architectural level via a multi-branch scheme or revealed the underlying correlation between them by unidirectional feature fusion, which disregards the explicit semantic and co-occurrence relations among objects and background. Inspired by the fact that context information is critical to recognize and localize the objects, and inclusive object details are significant to parse the background scene, we thus investigate on explicitly modeling the correlations between object and background to achieve a holistic understanding of an image in the panoptic segmentation task. We introduce a Bidirectional Graph Reasoning Network (BGRNet), which incorporates graph structure into the conventional panoptic segmentation network to mine the intra-modular and inter-modular relations within and between foreground things and background stuff classes. In particular, BGRNet first constructs image-specific graphs in both instance and semantic segmentation branches that enable flexible reasoning at the proposal level and class level, respectively. To establish the correlations between separate branches and fully leverage the complementary relations between things and stuff, we propose a Bidirectional Graph Connection Module to diffuse information across branches in a learnable fashion. Experimental results demonstrate the superiority of our BGRNet that achieves the new state-of-the-art performance on challenging COCO and ADE20K panoptic segmentation benchmarks.

Automatically evaluating dialogue coherence is a challenging but high-demand ability for developing high-quality open-domain dialogue systems.However, current evaluation metrics consider only surface features or utterancelevel semantics, without explicitly considering the fine-grained topic transition dynamics of dialogue flows.Here, we first consider that the graph structure constituted with topics in a dialogue can accurately depict the underlying communication logic, which is a more natural way to produce persuasive metrics.Capitalized on the topic-level dialogue graph, we propose a new evaluation metric GRADE, which stands for Graph-enhanced Representations for Automatic Dialogue Evaluation.Specifically, GRADE incorporates both coarsegrained utterance-level contextualized representations and fine-grained topic-level graph representations to evaluate dialogue coherence.The graph representations are obtained by reasoning over topic-level dialogue graphs enhanced with the evidence from a commonsense graph, including k-hop neighboring representations and hop-attention weights.Experimental results show that our GRADE significantly outperforms other state-of-the-art metrics on measuring diverse dialogue models in terms of the Pearson and Spearman correlations with human judgements.Besides, we release a new large-scale human evaluation benchmark to facilitate future research on automatic metrics.

A practical automatic textual math word problems (MWPs) solver should be able to solve various textual MWPs while most existing works only focused on one-unknown linear MWPs.Herein, we propose a simple but efficient method called Universal Expression Tree (UET) to make the first attempt to represent the equations of various MWPs uniformly.Then a semantically-aligned universal tree-structured solver (SAU-Solver) based on an encoder-decoder framework is proposed to resolve multiple types of MWPs in a unified model, benefiting from our UET representation.Our SAU-Solver generates a universal expression tree explicitly by deciding which symbol to generate according to the generated symbols' semantic meanings like human solving MWPs.Besides, our SAU-Solver also includes a novel subtree-level semanticallyaligned regularization to further enforce the semantic constraints and rationality of the generated expression tree by aligning with the contextual information.Finally, to validate the universality of our solver and extend the research boundary of MWPs, we introduce a new challenging Hybrid Math Word Problems dataset (HMWP), consisting of three types of MWPs.Experimental results on several MWPs datasets show that our model can solve universal types of MWPs and outperforms several state-of-the-art models 1 .

Similar to language, understanding an image can be considered as a hierarchical decomposition process from scenes to objects, parts, pixels, and the corresponding spatial/contextual relations. However, the existing convolutional networks concentrate on stacking redundant convolutional layers with a large number of kernels in a hierarchical organization to implicitly approximate this decomposition. This may limit the network to learn the semantic information conveyed in the internal feature maps that may reveal minor yet crucial differences for visual understanding. Attempting to tackle this problem, this paper proposes a simple yet effective tree convolution (TreeConv) operation for deep neural networks. Specifically, inspired by the image grammar techniques[73] that serve as a unified framework of object representation, learning, and recognition, our TreeConv designs a generative image grammar, i.e., tree generation rule, to parse the hierarchy of internal feature maps by generating tree structures and implicitly learning the specific visual grammars for each object category. Extensive experiments on a variety of benchmarks, i.e., classification (ImageNet / CIFAR), detection & segmentation (COCO 2017), and person re-identification (CUHK03), demonstrate the superiority of our TreeConv in both boosting the accuracy and reducing the computational cost. The source code will be available at: https://github.com/wanggrun/TreeConv.

Data inconsistency and bias are inevitable among different facial expression recognition (FER) datasets due to subjective annotating process and different collecting conditions. Recent works resort to adversarial mechanisms that learn domain-invariant features to mitigate domain shift. However, most of these works focus on holistic feature adaptation, and they ignore local features that are more transferable across different datasets. Moreover, local features carry more detailed and discriminative content for expression recognition, and thus integrating local features may enable fine-grained adaptation. In this work, we propose a novel Adversarial Graph Representation Adaptation (AGRA) framework that unifies graph representation propagation with adversarial learning for cross-domain holistic-local feature co-adaptation. To achieve this, we first build a graph to correlate holistic and local regions within each domain and another graph to correlate these regions across different domains. Then, we learn the per-class statistical distribution of each domain and extract holistic-local features from the input image to initialize the corresponding graph nodes. Finally, we introduce two stacked graph convolution networks to propagate holistic-local feature within each domain to explore their interaction and across different domains for holistic-local feature co-adaptation. In this way, the AGRA framework can adaptively learn fine-grained domain-invariant features and thus facilitate cross-domain expression recognition. We conduct extensive and fair experiments on several popular benchmarks and show that the proposed AGRA framework achieves superior performance over previous state-of-the-art methods.

Temporal grounding of natural language in untrimmed videos is a fundamental yet challenging multimedia task facilitating cross-media visual content retrieval. We focus on the weakly supervised setting of this task that merely accesses to coarse video-level language description annotation without temporal boundary, which is more consistent with reality as such weak labels are more readily available in practice. In this paper, we propose a Boundary Adaptive Refinement (BAR) framework that resorts to reinforcement learning (RL) to guide the process of progressively refining the temporal boundary. To the best of our knowledge, we offer the first attempt to extend RL to temporal localization task with weak supervision. As it is non-trivial to obtain a straightforward reward function in the absence of pairwise granular boundary-query annotations, a cross-modal alignment evaluator is crafted to measure the alignment degree of segment-query pair to provide tailor-designed rewards. This refinement scheme completely abandons traditional sliding window based solution pattern and contributes to acquiring more efficient, boundary-flexible and content-aware grounding results. Extensive experiments on two public benchmarks Charades-STA and ActivityNet demonstrate that BAR outperforms the state-of-the-art weakly-supervised method and even beats some competitive fully-supervised ones.

Crowd counting is an application-oriented task and its inference efficiency is crucial for real-world applications. However, most previous works relied on heavy backbone networks and required prohibitive run-time consumption, which would seriously restrict their deployment scopes and cause poor scalability. To liberate these crowd counting models, we propose a novel Structured Knowledge Transfer (SKT) framework, which fully exploits the structured knowledge of a well-trained teacher network to generate a lightweight but still highly effective student network. Specifically, it is integrated with two complementary transfer modules, including an Intra-Layer Pattern Transfer which sequentially distills the knowledge embedded in layer-wise features of the teacher network to guide feature learning of the student network and an Inter-Layer Relation Transfer which densely distills the cross-layer correlation knowledge of the teacher to regularize the student's feature evolution. Consequently, our student network can derive the layer-wise and cross-layer knowledge from the teacher network to learn compact yet effective features. Extensive evaluations on three benchmarks well demonstrate the effectiveness of our SKT for extensive crowd counting models. In particular, only using around $6%$ of the parameters and computation cost of original models, our distilled VGG-based models obtain at least 6.5× speed-up on an Nvidia 1080 GPU and even achieve state-of-the-art performance. Our code and models are available at https://github.com/HCPLab-SYSU/SKT.

With the development of deep learning, skeleton-based action recognition has achieved great progress in recent years. However, most of the current works focus on extracting more informative spatial representations of the human body, but haven't made full use of the temporal dependencies already contained in the sequence of human action. To this end, we propose a novel transformer-based model called Motion-Transformer to sufficiently capture the temporal dependencies via self-supervised pre-training on the sequence of human action. Besides, we propose to predict the motion flow of human skeletons for better learning the temporal dependencies in sequence. The pre-trained model is then fine-tuned on the task of action recognition. Experimental results on the large scale NTU RGB+D dataset shows our model is effective in modeling temporal relation, and the flow prediction pre-training is beneficial to expose the inherent dependencies in time dimensional. With this pre-training and fine-tuning paradigm, our final model outperforms previous state-of-the-art methods.

Panoptic segmentation is posed as a new popular test-bed for the state-of-the-art holistic scene understanding methods with the requirement of simultaneously segmenting both foreground things and background stuff. The state-of-the-art panoptic segmentation network exhibits high structural complexity in different network components, i.e. backbone, proposal-based foreground branch, segmentation-based background branch, and feature fusion module across branches, which heavily relies on expert knowledge and tedious trials. In this work, we propose an efficient, cooperative and highly automated framework to simultaneously search for all main components including backbone, segmentation branches, and feature fusion module in a unified panoptic segmentation pipeline based on the prevailing one-shot Network Architecture Search (NAS) paradigm. Notably, we extend the common single-task NAS into the multi-component scenario by taking the advantages of the newly proposed intra-modular search space and problem-oriented inter-modular search space, which helps us to obtain an optimal network architecture that not only performs well in both instance segmentation and semantic segmentation tasks but also be aware of the reciprocal relations between foreground things and background stuff classes. To relieve the vast computation burden incurred by applying NAS to complicated network architectures, we present a novel path-priority greedy search policy to find a robust, transferrable architecture with significantly reduced searching overhead. Our searched architecture, namely Auto-Panoptic, achieves the new state-of-the-art on the challenging COCO and ADE20K benchmarks. Moreover, extensive experiments are conducted to demonstrate the effectiveness of path-priority policy and transferability of Auto-Panoptic across different datasets.

Face hallucination is a domain-specific super-resolution problem that aims to generate a high-resolution (HR) face image from a low-resolution (LR) input. In contrast to the existing patch-wise super-resolution models that divide a face image into regular patches and independently apply LR to HR mapping to each patch, we implement deep reinforcement learning and develop a novel attention-aware face hallucination (Attention-FH) framework, which recurrently learns to attend a sequence of patches and performs facial part enhancement by fully exploiting the global interdependency of the image. Specifically, our proposed framework incorporates two components: a recurrent policy network for dynamically specifying a new attended region at each time step based on the status of the super-resolved image and the past attended region sequence, and a local enhancement network for selected patch hallucination and global state updating. The Attention-FH model jointly learns the recurrent policy network and local enhancement network through maximizing a long-term reward that reflects the hallucination result with respect to the whole HR image. Extensive experiments demonstrate that our Attention-FH significantly outperforms the state-of-the-art methods on in-the-wild face images with large pose and illumination variations.

Driven by recent computer vision and robotic applications, recovering 3D human poses has become increasingly important and attracted growing interests. In fact, completing this task is quite challenging due to the diverse appearances, viewpoints, occlusions and inherently geometric ambiguities inside monocular images. Most of the existing methods focus on designing some elaborate priors /constraints to directly regress 3D human poses based on the corresponding 2D human pose-aware features or 2D pose predictions. However, due to the insufficient 3D pose data for training and the domain gap between 2D space and 3D space, these methods have limited scalabilities for all practical scenarios (e.g., outdoor scene). Attempt to address this issue, this paper proposes a simple yet effective self-supervised correction mechanism to learn all intrinsic structures of human poses from abundant images. Specifically, the proposed mechanism involves two dual learning tasks, i.e., the 2D-to-3D pose transformation and 3D-to-2D pose projection, to serve as a bridge between 3D and 2D human poses in a type of "free" self-supervision for accurate 3D human pose estimation. The 2D-to-3D pose implies to sequentially regress intermediate 3D poses by transforming the pose representation from the 2D domain to the 3D domain under the sequence-dependent temporal context, while the 3D-to-2D pose projection contributes to refining the intermediate 3D poses by maintaining geometric consistency between the 2D projections of 3D poses and the estimated 2D poses. Therefore, these two dual learning tasks enable our model to adaptively learn from 3D human pose data and external large-scale 2D human pose data. We further apply our self-supervised correction mechanism to develop a 3D human pose machine, which jointly integrates the 2D spatial relationship, temporal smoothness of predictions and 3D geometric knowledge. Extensive evaluations on the Human3.6M and HumanEva-I benchmarks demonstrate the superior performance and efficiency of our framework over all the compared competing methods.

Different from traditional image super-resolution task, real image super-resolution(Real-SR) focus on the relationship between real-world high-resolution(HR) and low-resolution(LR) image. Most of the traditional image SR obtains the LR sample by applying a fixed down-sampling operator. Real-SR obtains the LR and HR image pair by incorporating different quality optical sensors. Generally, Real-SR has more challenges as well as broader application scenarios. Previous image SR methods fail to exhibit similar performance on Real-SR as the image data is not aligned inherently. In this article, we propose a Dual-path Dynamic Enhancement Network(DDet) for Real-SR, which addresses the cross-camera image mapping by realizing a dual-way dynamic sub-pixel weighted aggregation and refinement. Unlike conventional methods which stack up massive convolutional blocks for feature representation, we introduce a content-aware framework to study non-inherently aligned image pair in image SR issue. First, we use a content-adaptive component to exhibit the Multi-scale Dynamic Attention(MDA). Second, we incorporate a long-term skip connection with a Coupled Detail Manipulation(CDM) to perform collaborative compensation and manipulation. The above dual-path model is joint into a unified model and works collaboratively. Extensive experiments on the challenging benchmarks demonstrate the superiority of our model.

Model drifting is one of the knotty problems that seriously restricts the accuracy of discriminative trackers in visual tracking. Most existing works usually focus on improving the robustness of the target appearance model. However, they are prone to suffer from model drifting due to the inappropriate model updates during the tracking-by-detection. In this paper, we propose a novel update-pacing framework to suppress the occurrence of model drifting in visual tracking. Specifically, the proposed framework first initializes an ensemble of trackers, each of which updates the model in a different update interval. Once the forward tracking trajectory of each tracker is determined, the backward trajectory will also be generated by the current model to measure the difference with the forward one, and the tracker with the smallest deviation score will be selected as the most robust tracker for the remaining tracking. By performing such self-examination on trajectory pairs, the framework can effectively preserve the temporal context consistency of sequential frames to avoid learning corrupted information. To further improve the performance of the proposed method, a multi-feature extension framework is also proposed to incorporate multiple features into the ensemble of the trackers. The extensive experimental results obtained on large-scale object tracking benchmarks demonstrate that the proposed framework significantly increases the accuracy and robustness of the underlying base trackers, such as DSST, Struck, KCF, and CT, and achieves superior performance compared with the state-of-the-art methods without using deep models.

The field of computer vision has witnessed phenomenal progress in recent years partially due to the development of deep convolutional neural networks. However, deep learning models are notoriously sensitive to adversarial examples which are synthesized by adding quasi-perceptible noises on real images. Some existing defense methods require to re-train attacked target networks and augment the train set via known adversarial attacks, which is inefficient and might be unpromising with unknown attack types. To overcome the above issues, we propose a portable defense method, online alternate generator, which does not need to access or modify the parameters of the target networks. The proposed method works by online synthesizing another image from scratch for an input image, instead of removing or destroying adversarial noises. To avoid pretrained parameters exploited by attackers, we alternately update the generator and the synthesized image at the inference stage. Experimental results demonstrate that the proposed defensive scheme and method outperforms a series of state-of-the-art defending models against gray-box adversarial attacks.

This article presents a semisupervised multilabel fully convolutional network (FCN) for hierarchical object parsing of images. We consider each object part (e.g., eye and head) as a class label and learn to assign every image pixel to multiple coherent part labels. Different from previous methods that consider part labels as independent classes, our method explicitly models the internal relationships between object parts, e.g., that a pixel highly scored for eyes should be highly scored for heads as well. Such relationships directly reflect the structure of the semantic space and thus should be respected while learning the deep representation. We achieve this objective by introducing a multilabel softmax loss function over both labeled and unlabeled images and regularizing it with two pairwise ranking constraints. The first constraint is based on a manifold assumption that image pixels being visually and spatially close to each other should be collaboratively classified as the same part label. The other constraint is used to enforce that no pixel receives significant scores from more than one label that are semantically conflicting with each other. The proposed loss function is differentiable with respect to network parameters and hence can be optimized by standard stochastic gradient methods. We evaluate the proposed method on two public image data sets for hierarchical object parsing and compare it with the alternative parsing methods. Extensive comparisons showed that our method can achieve state-of-the-art performance while using 50% less labeled training samples than the alternatives.

Facial action unit (AU) recognition is a crucial task for facial expressions analysis and has attracted extensive attention in the field of artificial intelligence and computer vision. Existing works have either focused on designing or learning complex regional feature representations, or delved into various types of AU relationship modeling. Albeit with varying degrees of progress, it is still arduous for existing methods to handle complex situations. In this paper, we investigate how to integrate the semantic relationship propagation between AUs in a deep neural network framework to enhance the feature representation of facial regions, and propose an AU semantic relationship embedded representation learning (SRERL) framework. Specifically, by analyzing the symbiosis and mutual exclusion of AUs in various facial expressions, we organize the facial AUs in the form of structured knowledge-graph and integrate a Gated Graph Neural Network (GGNN) in a multi-scale CNN framework to propagate node information through the graph for generating enhanced AU representation. As the learned feature involves both the appearance characteristics and the AU relationship reasoning, the proposed model is more robust and can cope with more challenging cases, e.g., illumination change and partial occlusion. Extensive experiments on the two public benchmarks demonstrate that our method outperforms the previous work and achieves state of the art performance.

To force convolutional networks to explore more discriminative evidence throughout spatial regions, this paper presents a novel CamDrop to improve the conventional dropout in two aspects. First, by considering the intensity of class activation mapping (CAM) all around, CamDrop selectively abandons some specific spatial regions in predominating visual patterns at each iteration. In many classification tasks, CamDrop demonstrates its effectiveness and achieves considerable improvements on robust predictions for adversarial examples. Second, although dropout is a widely adopted technique that has been applied to regularize large models, the improvement in performance always attributes to better preventing DNN from overfitting. Here we give a new explanation of dropout from the perspective of optimization that it makes the upper bound of the magnitude of gradients much tighter, which leads to a more stable behavior of the gradients and effectively avoids neurons falling into the saturation region of the nonlinear activation, even when using high learning rates. Extensive experiments have been performed to prove the above two strengths of CamDrop.

Recent studies have shown remarkable advances in 3D human pose estimation from monocular images, with the help of large-scale in-door 3D datasets and sophisticated network architectures. However, the generalizability to different environments remains an elusive goal. In this work, we propose a geometry-aware 3D representation for the human pose to address this limitation by using multiple views in a simple auto-encoder model at the training stage and only 2D keypoint information as supervision. A view synthesis framework is proposed to learn the shared 3D representation between viewpoints with synthesizing the human pose from one viewpoint to the other one. Instead of performing a direct transfer in the raw image-level, we propose a skeleton-based encoder-decoder mechanism to distil only pose-related representation in the latent space. A learning-based representation consistency constraint is further introduced to facilitate the robustness of latent 3D representation. Since the learnt representation encodes 3D geometry information, mapping it to 3D pose will be much easier than conventional frameworks that use an image or 2D coordinates as the input of 3D pose estimator. We demonstrate our approach on the task of 3D human pose estimation. Comprehensive experiments on three popular benchmarks show that our model can significantly improve the performance of state-of-the-art methods with simply injecting the representation as a robust 3D prior.

Current neural networks for 3D object recognition are vulnerable to 3D rotation. Existing works mostly rely on massive amounts of rotation-augmented data to alleviate the problem, which lacks solid guarantee of the 3D rotation invariance. In this paper, we address the issue by introducing a novel point cloud representation that can be mathematically proved rigorously rotation-invariant, i.e., identical point clouds in different orientations are unified as a unique and consistent representation. Moreover, the proposed representation is conditional information-lossless, because it retains all necessary information of point cloud except for orientation information. In addition, the proposed representation is complementary with existing network architectures for point cloud and fundamentally improves their robustness against rotation transformation. Finally, we propose a deep hierarchical cluster network called ClusterNet to better adapt to the proposed representation. We employ hierarchical clustering to explore and exploit the geometric structure of point cloud, which is embedded in a hierarchical structure tree. Extensive experimental results have shown that our proposed method greatly outperforms the state-of-the-arts in rotation robustness on rotation-augmented 3D object classification benchmarks.

To understand a scene in depth not only involves locating/recognizing individual objects, but also requires to infer the relationships and interactions among them. However, since the distribution of real-world relationships is seriously unbalanced, existing methods perform quite poorly for the less frequent relationships. In this work, we find that the statistical correlations between object pairs and their relationships can effectively regularize semantic space and make prediction less ambiguous, and thus well address the unbalanced distribution issue. To achieve this, we incorporate these statistical correlations into deep neural networks to facilitate scene graph generation by developing a Knowledge-Embedded Routing Network. More specifically, we show that the statistical correlations between objects appearing in images and their relationships, can be explicitly represented by a structured knowledge graph, and a routing mechanism is learned to propagate messages through the graph to explore their interactions. Extensive experiments on the large-scale Visual Genome dataset demonstrate the superiority of the proposed method over current state-of-the-art competitors.

(Unsupervised) Domain Adaptation (DA) seeks for classifying target instances when solely provided with source labeled and target unlabeled examples for training. Learning domain-invariant features helps to achieve this goal, whereas it underpins unlabeled samples drawn from a single or multiple explicit target domains (Multi-target DA). In this paper, we consider a more realistic transfer scenario: our target domain is comprised of multiple sub-targets implicitly blended with each other so that learners could not identify which sub-target each unlabeled sample belongs to. This Blending-target Domain Adaptation (BTDA) scenario commonly appears in practice and threatens the validities of existing DA algorithms, due to the presence of domain gaps and categorical misalignments among these hidden sub-targets. To reap the transfer performance gains in this new scenario, we propose Adversarial Meta-Adaptation Network (AMEAN). AMEAN entails two adversarial transfer learning processes. The first is a conventional adversarial transfer to bridge our source and mixed target domains. To circumvent the intra-target category misalignment, the second process presents as ``learning to adapt'': It deploys an unsupervised meta-learner receiving target data and their ongoing feature-learning feedbacks, to discover target clusters as our ``meta-sub-target'' domains. This meta-sub-targets auto-design our meta-sub-target adaptation loss, which is capable to progressively eliminate the implicit category mismatching in our mixed target. We evaluate AMEAN and a variety of DA algorithms in three benchmarks under the BTDA setup. Empirical results show that BTDA is a quite challenging transfer setup for most existing DA algorithms, yet AMEAN significantly outperforms these state-of-the-art baselines and effectively restrains the negative transfer effects in BTDA.

Prior highly-tuned human parsing models tend to fit towards each dataset in a specific domain or with discrepant label granularity, and can hardly be adapted to other human parsing tasks without extensive re-training. In this paper, we aim to learn a single universal human parsing model that can tackle all kinds of human parsing needs by unifying label annotations from different domains or at various levels of granularity. This poses many fundamental learning challenges, e.g. discovering underlying semantic structures among different label granularity, performing proper transfer learning across different image domains, and identifying and utilizing label redundancies across related tasks. To address these challenges, we propose a new universal human parsing agent, named “Graphonomy”, which incorporates hierarchical graph transfer learning upon the conventional parsing network to encode the underlying label semantic structures and propagate relevant semantic information. In particular, Graphonomy first learns and propagates compact high-level graph representation among the labels within one dataset via Intra-Graph Reasoning, and then transfers semantic information across multiple datasets via Inter-Graph Transfer. Various graph transfer dependencies (e.g., similarity, linguistic knowledge) between different datasets are analyzed and encoded to enhance graph transfer capability. By distilling universal semantic graph representation to each specific task, Graphonomy is able to predict all levels ofparsing labels in one system without piling up the complexity. Experimental results show Graphonomy effectively achieves the state-of-the-art results on three human parsing benchmarks as well as advantageous universal human parsing performance.

Joint filtering mainly uses an additional guidance image as a prior and transfers its structures to the target image in the filtering process. Different from existing algorithms that rely on locally linear models or hand-designed objective functions to extract the structural information from the guidance image, we propose a new joint filter based on a spatially variant linear representation model (SVLRM), where the target image is linearly represented by the guidance image. However, the SVLRM leads to a highly ill-posed problem. To estimate the linear representation coefficients, we develop an effective algorithm based on a deep convolutional neural network (CNN). The proposed deep CNN (constrained by the SVLRM) is able to estimate the spatially variant linear representation coefficients which are able to model the structural information of both the guidance and input images. We show that the proposed algorithm can be effectively applied to a variety of applications, including depth/RGB image upsampling and restoration, flash/no-flash image deblurring, natural image denoising, scale-aware filtering, etc. Extensive experimental results demonstrate that the proposed algorithm performs favorably against state-of-the-art methods that have been specially designed for each task.

This paper presents a novel adaptively connected neural network (ACNet) to improve the traditional convolutional neural networks (CNNs) {in} two aspects. First, ACNet employs a flexible way to switch global and local inference in processing the internal feature representations by adaptively determining the connection status among the feature nodes (e.g., pixels of the feature maps). Note that in a computer vision domain, a node refers to a pixel of a feature map, while in the graph domain, a node denotes a graph node. We can show that existing CNNs, the classical multilayer perceptron (MLP), and the recently proposed non-local network (NLN) are all special cases of ACNet. Second, ACNet is also capable of handling non-Euclidean data. Extensive experimental analyses on a variety of benchmarks (i.e., ImageNet-1k classification, COCO 2017 detection and segmentation, CUHK03 person re-identification, CIFAR analysis, and Cora document categorization) demonstrate that ACNet cannot only achieve state-of-the-art performance but also overcome the limitation of the conventional MLP and CNN. The code is available at https://github.com/wanggrun/Adaptively-Connected-Neural-Networks.

In this paper, we address the large-scale object detection problem with thousands of categories, which poses severe challenges due to long-tail data distributions, heavy occlusions, and class ambiguities. However, the dominant object detection paradigm is limited by treating each object region separately without considering crucial semantic dependencies among objects. In this work, we introduce a novel Reasoning-RCNN to endow any detection networks the capability of adaptive global reasoning over all object regions by exploiting diverse human commonsense knowledge. Instead of only propagating the visual features on the image directly, we evolve the high-level semantic representations of all categories globally to avoid distracted or poor visual features in the image. Specifically, built on feature representations of basic detection network, the proposed network first generates a global semantic pool by collecting the weights of previous classification layer for each category, and then adaptively enhances each object features via attending different semantic contexts in the global semantic pool. Rather than propagating information from all semantic information that may be noisy, our adaptive global reasoning automatically discovers most relative categories for feature evolving. Our Reasoning-RCNN is light-weight and flexible enough to enhance any detection backbone networks, and extensible for integrating any knowledge resources. Solid experiments on object detection benchmarks show the superiority of our Reasoning-RCNN, e.g. achieving around 16% improvement on VisualGenome, 37% on ADE in terms of mAP and 15% improvement on COCO.

Recognizing multiple labels of images is a practical and challenging task, and significant progress has been made by searching semantic-aware regions and modeling label dependency. However, current methods cannot locate the semantic regions accurately due to the lack of part-level supervision or semantic guidance. Moreover, they cannot fully explore the mutual interactions among the semantic regions and do not explicitly model the label co-occurrence. To address these issues, we propose a Semantic-Specific Graph Representation Learning (SSGRL) framework that consists of two crucial modules: 1) a semantic decoupling module that incorporates category semantics to guide learning semantic-specific representations and 2) a semantic interaction module that correlates these representations with a graph built on the statistical label co-occurrence and explores their interactions via a graph propagation mechanism. Extensive experiments on public benchmarks show that our SSGRL framework outperforms current state-of-the-art methods by a sizable margin, e.g. with an mAP improvement of 2.5%, 2.6%, 6.7%, and 3.1% on the PASCAL VOC 2007 & 2012, Microsoft-COCO and Visual Genome benchmarks, respectively. Our codes and models are available at https://github.com/HCPLab-SYSU/SSGRL.

Matching clothing images from customers and online shopping stores has rich applications in E-commerce. Existing algorithms encoded an image as a global feature vector and performed retrieval with the global representation. However, discriminative local information on clothes are submerged in this global representation, resulting in sub-optimal performance. To address this issue, we propose a novel Graph Reasoning Network (GRNet) on a Similarity Pyramid, which learns similarities between a query and a gallery cloth by using both global and local representations in multiple scales. The similarity pyramid is represented by a Graph of similarity, where nodes represent similarities between clothing components at different scales, and the final matching score is obtained by message passing along edges. In GRNet, graph reasoning is solved by training a graph convolutional network, enabling to align salient clothing components to improve clothing retrieval. To facilitate future researches, we introduce a new benchmark FindFashion, containing rich annotations of bounding boxes, views, occlusions, and cropping. Extensive experiments show that GRNet obtains new state-of-the-art results on two challenging benchmarks, e.g. pushing the top-1, top-20, and top-50 accuracies on DeepFashion to 26%, 64%, and 75% (i.e. 4%, 10%, and 10% absolute improvements), outperforming competitors with large margins. On FindFashion, GRNet achieves considerable improvements on all empirical settings.

Automatic estimation of the number of people in unconstrained crowded scenes is a challenging task and one major difficulty stems from the huge scale variation of people. In this paper, we propose a novel Deep Structured Scale Integration Network (DSSINet) for crowd counting, which addresses the scale variation of people by using structured feature representation learning and hierarchically structured loss function optimization. Unlike conventional methods which directly fuse multiple features with weighted average or concatenation, we first introduce a Structured Feature Enhancement Module based on conditional random fields (CRFs) to refine multiscale features mutually with a message passing mechanism. Specifically, each scale-specific feature is considered as a continuous random variable and passes complementary information to refine the features at other scales. Second, we utilize a Dilated Multiscale Structural Similarity loss to enforce our DSSINet to learn the local correlation of people's scales within regions of various size, thus yielding high-quality density maps. Extensive experiments on four challenging benchmarks well demonstrate the effectiveness of our method. In particular, our DSSINet achieves improvements of 9.5% error reduction on Shanghaitech dataset and 24.9% on UCF-QNRF dataset against the state-of-the-art methods.

Domain adaptation enables the learner to safely generalize into novel environments by mitigating domain shifts across distributions. Previous works may not effectively uncover the underlying reasons that would lead to the drastic model degradation on the target task. In this paper, we empirically reveal that the erratic discrimination of the target domain mainly stems from its much smaller feature norms with respect to that of the source domain. To this end, we propose a novel parameter-free Adaptive Feature Norm approach. We demonstrate that progressively adapting the feature norms of the two domains to a large range of values can result in significant transfer gains, implying that those task-specific features with larger norms are more transferable. Our method successfully unifies the computation of both standard and partial domain adaptation with more robustness against the negative transfer issue. Without bells and whistles but a few lines of code, our method substantially lifts the performance on the target task and exceeds state-of-the-arts by a large margin (11.5% on Office-Home and 17.1% on VisDA2017). We hope our simple yet effective approach will shed some light on the future research of transfer learning. Code is available at https://github.com/jihanyang/AFN.

Resembling the rapid learning capability of human, low-shot learning empowers vision systems to understand new concepts by training with few samples. Leading approaches derived from meta-learning on images with a single visual object. Obfuscated by a complex background and multiple objects in one image, they are hard to promote the research of low-shot object detection/segmentation. In this work, we present aflexible and general methodology to achieve these tasks. Our work extends Faster /Mask R-CNN by proposing meta-learning over RoI (Region-of-Interest) features instead of a full image feature. This simple spirit disentangles multi-object information merged with the background, without bells and whistles, enabling Faster /Mask R-CNN turn into a meta-learner to achieve the tasks. Specifically, we introduce a Predictor-head Remodeling Network (PRN) that shares its main backbone with Faster /Mask R-CNN. PRN receives images containing low-shot objects with their bounding boxes or masks to infer their class attentive vectors. The vectors take channel-wise soft-attention on RoI features, remodeling those R-CNN predictor heads to detect or segment the objects consistent with the classes these vectors represent. In our experiments, Meta R-CNN yields the new state of the art in low-shot object detection and improves low-shot object segmentation byMaskR-CNN.Code: https://yanxp.github.io/metarcnn.html.

Deep learning-based video salient object detection has recently achieved great success with its performance significantly outperforming any other unsupervised methods. However, existing data-driven approaches heavily rely on a large quantity of pixel-wise annotated video frames to deliver such promising results. In this paper, we address the semi-supervised video salient object detection task using pseudo-labels. Specifically, we present an effective video saliency detector that consists of a spatial refinement network and a spatiotemporal module. Based on the same refinement network and motion information in terms of optical flow, we further propose a novel method for generating pixel-level pseudo-labels from sparsely annotated frames. By utilizing the generated pseudo-labels together with a part of manual annotations, our video saliency detector learns spatial and temporal cues for both contrast inference and coherence enhancement, thus producing accurate saliency maps. Experimental results demonstrate that our proposed semi-supervised method even greatly outperforms all the state-of-the-art fully supervised methods across three public benchmarks of VOS, DAVIS, and FBMS.

Taxi demand prediction has recently attracted increasing research interest due to its huge potential application in large-scale intelligent transportation systems. However, most of the previous methods only considered the taxi demand prediction in origin regions, while ignoring the modeling of the specific situation of the destination passengers. In this paper, we present a more challenging task, called taxi origin-destination demand prediction, which aims at predicting the taxi demand between all origin-destination (OD) pairs in a future time interval. Its main challenges lie in how to effectively capture the diverse contextual information to learn the demand patterns. We address this problem with a novel Contextualized Spatial-Temporal Network (CSTN), which can effectively capture various context of taxi demand into a unified framework. Specifically, the proposed network consists of three components for the modeling of local spatial context (LSC), temporal evolution context (TEC) and global correlation context (GCC) respectively. Extensive experiments and evaluations on a large-scale dataset well demonstrate the significant superiority of our CSTN over other compared methods of taxi origin-destination demand prediction.

Crowd counting, aiming at estimating the total number of people in unconstrained crowded scenes, has increasingly received attention. But it is greatly challenged by the huge variation in people scale. In this paper, we propose a novel Multi-View Scale Aggregation Network (MVSAN), which handle the scale variation from feature, input and criterion view comprehensively. Firstly, we design a simple but effective Multi-Scale Feature Encoder, which exploits dilated convolution layers with various dilation rates to improve the representation ability and scale diversity of features. Secondly, we feed multiple scales of input images into networks to generate high-quality density maps in a coarse-to-fine manner. Finally, we propose a Multi-Scale Structural Similarity loss to force our networks to learn the local correlation of density maps. Extensive experiments on two standard benchmarks show that the proposed method can generate high-quality crowd density map and accurate count estimation, outperforming the state-of-the-art methods with a large margin.

Current methods for image captioning tend to generate sentences that are generally overly rigid and composed of some most frequent words/phrases, leading to inaccurate and indistinguishable descriptions. This is primarily due to the uneven word distribution of the ground truth captions that encourages to generate high frequent words/phrases while suppressing the less frequent but more concrete ones. In this work, we propose a new Content Sensitive and Global Discriminative objective, which is formulated as two constraints on top of a reference model to facilitate generating concrete and discriminative image captions. More specifically, the content sensitive constraint is designed to place greater focus on the less frequent and more concrete words/phrases, thus facilitating the generation of sentences that better describe visual details of the given images. To further improve the discriminability, the global discriminative constraint is designed to pull the generated sentence to better discern the corresponding image from others. We evaluate the proposed method on the widely used MS-COCO dataset, where it achieves superior performance over existing competing methods. We also conduct self-retrieval experiments to demonstrate the discriminability of the proposed method.

A broad range of cross-$m$-domain generation researches boil down to matching a joint distribution by deep generative models (DGMs). Hitherto algorithms excel in pairwise domains while as $m$ increases, remain struggling to scale themselves to fit a joint distribution. In this paper, we propose a domain-scalable DGM, i.e., MMI-ALI for $m$-domain joint distribution matching. As an $m$-domain ensemble model of ALIs (Dumoulin et al., 2016), MMI-ALI is adversarially trained with maximizing Multivariate Mutual Information (MMI) w.r.t. joint variables of each pair of domains and their shared feature. The negative MMIs are upper bounded by a series of feasible losses provably leading to matching $m$-domain joint distributions. MMI-ALI linearly scales as $m$ increases and thus, strikes a right balance between efficacy and scalability. We evaluate MMI-ALI in diverse challenging $m$-domain scenarios and verify its superiority.

Salient object detection, which aims at localizing the attention-aware visual objects, is the indispensable technology for intelligent robots to understand and interact with the complicated environments. Existing salient object detection approaches mainly focus on the optimization of detection performance, while ignoring the considerations for computational resource consumption and algorithm efficiency. Contrarily, we build a superior lightweight network architecture to simultaneously improve performance on both accuracy and efficiency for salient object detection. Specifically, our proposed approach adopts the lightweight bottleneck as its primary building block to significantly reduce the number of parameters and to speed up the process of training and inference. In practice, the visual contrast is insufficiently discovered with the limitation of the small empirical receptive field of CNN. To alleviate this issue, we design a multi-scale convolution module to rapidly discover high-level visual contrast. Moreover, a lightweight refinement module is utilized to restore object saliency details with negligible extra cost. Extensive experiments on efficiency and accuracy trade-offs show that our model is more competitive than the state-of-the-art works on salient object detection task and has prominent potentials for robots applications in real time.

Existing visual tracking methods usually localize a target object with a bounding box, in which the performance of the foreground object trackers or detectors is often affected by the inclusion of background clutter. To handle this problem, we learn a patch-based graph representation for visual tracking. The tracked object is modeled by with a graph by taking a set of non-overlapping image patches as nodes, in which the weight of each node indicates how likely it belongs to the foreground and edges are weighted for indicating the appearance compatibility of two neighboring nodes. This graph is dynamically learned and applied in object tracking and model updating. During the tracking process, the proposed algorithm performs three main steps in each frame. First, the graph is initialized by assigning binary weights of some image patches to indicate the object and background patches according to the predicted bounding box. Second, the graph is optimized to refine the patch weights by using a novel alternating direction method of multipliers. Third, the object feature representation is updated by imposing the weights of patches on the extracted image features. The object location is predicted by maximizing the classification score in the structured support vector machine. Extensive experiments show that the proposed tracking algorithm performs well against the state-of-the-art methods on large-scale benchmark datasets.

Human parsing and pose estimation have recently received considerable interest due to their substantial application potentials. However, the existing datasets have limited numbers of images and annotations and lack a variety of human appearances and coverage of challenging cases in unconstrained environments. In this paper, we introduce a new benchmark named "Look into Person (LIP)" that provides a significant advancement in terms of scalability, diversity, and difficulty, which are crucial for future developments in human-centric analysis. This comprehensive dataset contains over 50,000 elaborately annotated images with 19 semantic part labels and 16 body joints, which are captured from a broad range of viewpoints, occlusions, and background complexities. Using these rich annotations, we perform detailed analyses of the leading human parsing and pose estimation approaches, thereby obtaining insights into the successes and failures of these methods. To further explore and take advantage of the semantic correlation of these two tasks, we propose a novel joint human parsing and pose estimation network to explore efficient context modeling, which can simultaneously predict parsing and pose with extremely high quality. Furthermore, we simplify the network to solve human parsing by exploring a novel self-supervised structure-sensitive learning approach, which imposes human pose structures into the parsing results without resorting to extra supervision. The datasets, code and models are available at http://www.sysu-hcp.net/lip/.

This paper investigates a fundamental problem of scene understanding: how to parse a scene image into a structured configuration (i.e., a semantic object hierarchy with object interaction relations). We propose a deep architecture consisting of two networks: i) a convolutional neural network (CNN) extracting the image representation for pixel-wise object labeling and ii) a recursive neural network (RsNN) discovering the hierarchical object structure and the inter-object relations. Rather than relying on elaborative annotations (e.g., manually labeled semantic maps and relations), we train our deep model in a weakly-supervised learning manner by leveraging the descriptive sentences of the training images. Specifically, we decompose each sentence into a semantic tree consisting of nouns and verb phrases, and apply these tree structures to discover the configurations of the training images. Once these scene configurations are determined, then the parameters of both the CNN and RsNN are updated accordingly by back propagation. The entire model training is accomplished through an Expectation-Maximization method. Extensive experiments show that our model is capable of producing meaningful scene configurations and achieving more favorable scene labeling results on two benchmarks (i.e., PASCAL VOC 2012 and SYSU-Scenes) compared with other state-of-the-art weakly-supervised deep learning methods. In particular, SYSU-Scenes contains more than 5,000 scene images with their semantic sentence descriptions, which is created by us for advancing research on scene parsing.

For visual tracking methods based on kernel support vector machines (SVMs), data sampling is usually adopted to reduce the computational cost in training. In addition, budgeting of support vectors is required for computational efficiency. Instead of sampling and budgeting, recently the circulant matrix formed by dense sampling of translated image patches has been utilized in kernel correlation filters for fast tracking. In this paper, we derive an equivalent formulation of a SVM model with the circulant matrix expression and present an efficient alternating optimization method for visual tracking. We incorporate the discrete Fourier transform with the proposed alternating optimization process, and pose the tracking problem as an iterative learning of support correlation filters (SCFs). In the fully-supervision setting, our SCF can find the globally optimal solution with real-time performance. For a given circulant data matrix with$n^2$samples of$n \times n$pixels, the computational complexity of the proposed algorithm is$O(n^2\; \log n)$whereas that of the standard SVM-based approaches is at least$O(n^4)$. In addition, we extend the SCF-based tracking algorithm with multi-channel features, kernel functions, and scale-adaptive approaches to further improve the tracking performance. Experimental results on a large benchmark dataset show that the proposed SCF-based algorithms perform favorably against the state-of-the-art tracking methods in terms of accuracy and speed.

In recent times, image inpainting has witnessed rapid progress due to the generative adversarial networks (GANs) that are able to synthesize realistic contents. However, most existing GAN-based methods for semantic inpainting apply an auto-encoder architecture with a fully connected layer, which cannot accurately maintain spatial information. In addition, the discriminator in existing GANs struggles to comprehend high-level semantics within the image context and yields semantically consistent content. Existing evaluation criteria are biased toward blurry results and cannot well characterize edge preservation and visual authenticity in the inpainting results. In this paper, we propose an improved GAN to overcome the aforementioned limitations. Our proposed GAN-based framework consists of a fully convolutional design for the generator which helps to better preserve spatial structures and a joint loss function with a revised perceptual loss to capture high-level semantics in the context. Furthermore, we also introduce two novel measures to better assess the quality of image inpainting results. The experimental results demonstrate that our method outperforms the state-of-the-art under a wide range of criteria.

Recognizing objects from simultaneously sensed photometric (RGB) and depth channels is a fundamental yet practical problem in many machine vision applications, such as robot grasping and autonomous driving. In this paper, we address this problem by developing a cross-modal attentional context (CMAC) learning framework, which enables the full exploitation of the context information from both RGB and depth data. Compared to existing RGB-D object detection frameworks, our approach has several appealing properties. First, it consists of an attention-based global context model for exploiting adaptive contextual information and incorporating this information into a region-based CNN (e.g., fast RCNN) framework to achieve improved object detection performance. Second, our CMAC framework further contains a fine-grained object part attention module to harness multiple discriminative object parts inside each possible object region for superior local feature representation. While greatly improving the accuracy of RGB-D object detection, the effective cross-modal information fusion as well as attentional context modeling in our proposed model provide an interpretable visualization scheme. Experimental results demonstrate that the proposed method significantly improves upon the state of the art on all public benchmarks.

Video person re-identification has attracted much attention in recent years. It aims to match image sequences of pedestrians from different camera views. Previous approaches usually improve this task from three aspects, including: 1) selecting more discriminative frames; 2) generating more informative temporal representations; and 3) developing more effective distance metrics. To address the above issues, we present a novel and practical deep architecture for video person re-identification termed self-and-collaborative attention network (SCAN), which adopts the video pairs as the input and outputs their matching scores. SCAN has several appealing properties. First, SCAN adopts a non-parametric attention mechanism to refine the intra-sequence and inter-sequence feature representation of videos and outputs self-and-collaborative feature representation for each video, making the discriminative frames aligned between the probe and gallery sequences. Second, beyond the existing models, a generalized pairwise similarity measurement is proposed to generate the similarity feature representation of video pair by calculating the Hadamard product of their self-representation difference and collaborative-representation difference. Thus, the matching result can be predicted by the binary classifier. Third, a dense clip segmentation strategy is also introduced to generate rich probe-gallery pairs to optimize the model. In the test phase, the final matching score of two videos is determined by averaging the scores of top-ranked clip-pairs. Extensive experiments demonstrate the effectiveness of SCAN, which outperforms the top-1 accuracies of the best-performing baselines on iLIDS-VID, PRID2011, and MARS datasets, respectively.

Taxi demand prediction has recently attracted increasing research interest due to its huge potential application in large-scale intelligent transportation systems. However, most of the previous methods only considered the taxi demand prediction in origin regions, but neglected the modeling of the specific situation of the destination passengers. We believe it is suboptimal to preallocate the taxi into each region-based solely on the taxi origin demand. In this paper, we present a challenging and worth-exploring task, called taxi origin-destination demand prediction, which aims at predicting the taxi demand between all-region pairs in a future time interval. Its main challenges come from how to effectively capture the diverse contextual information to learn the demand patterns. We address this problem with a novel contextualized spatial-temporal network (CSTN), which consists of three components for the modeling of local spatial context (LSC), temporal evolution context (TEC), and global correlation context (GCC), respectively. First, an LSC module utilizes two convolution neural networks to learn the local spatial dependencies of taxi, demand respectively, from the origin view and the destination view. Second, a TEC module incorporates the local spatial features of taxi demand and the meteorological information to a Convolutional Long Short-term Memory Network (ConvLSTM) for the analysis of taxi demand evolution. Finally, a GCC module is applied to model the correlation between all regions by computing a global correlation feature as a weighted sum of all regional features, with the weights being calculated as the similarity between the corresponding region pairs. The extensive experiments and evaluations on a large-scale dataset well demonstrate the superiority of our CSTN over other compared methods for the taxi origin-destination demand prediction.

This paper focuses on semantic task planning, that is, predicting a sequence of actions toward accomplishing a specific task under a certain scene, which is a new problem in computer vision research. The primary challenges are how to model the task-specific knowledge and how to integrate this knowledge into the learning procedure. In this paper, we propose training a recurrent long short-term memory (LSTM) network to address this problem, that is, taking a scene image (including prelocated objects) and the specified task as input and recurrently predicting action sequences. However, training such a network generally requires large numbers of annotated samples to cover the semantic space (e.g., diverse action decomposition and ordering). To overcome this issue, we introduce a knowledge and-or graph (AOG) for task description, which hierarchically represents a task as atomic actions. With this AOG representation, we can produce many valid samples (i.e., action sequences according to common sense) by training another auxiliary LSTM network with a small set of annotated samples. Furthermore, these generated samples (i.e., task-oriented action sequences) effectively facilitate training of the model for semantic task planning. In our experiments, we create a new dataset that contains diverse daily tasks and extensively evaluates the effectiveness of our approach.

Facial landmark localization plays a critical role in face recognition and analysis. In this paper, we propose a novel cascaded backbone-branches fully convolutional neural network (BB-FCN) for rapidly and accurately localizing facial landmarks in unconstrained and cluttered settings. Our proposed BB-FCN generates facial landmark response maps directly from raw images without any preprocessing. BB-FCN follows a coarse-to-fine cascaded pipeline, which consists of a backbone network to roughly detect the locations of all facial landmarks and one branch network for each type of detected landmark to further refine its location. Furthermore, to facilitate the facial landmark localization under unconstrained settings, we propose a large-scale benchmark named SYSU16K, which contains 16 000 faces with large variations in pose, expression, illumination, and resolution. Extensive experimental evaluations demonstrate that our proposed BB-FCN can significantly outperform the state of the art under both constrained (i.e., within detected facial regions only) and unconstrained settings. We further confirm that high-quality facial landmarks localized with our proposed network can also improve the precision and recall of face detection.

Though quite challenging, leveraging large-scale unlabeled or partially labeled data in learning systems (e.g., model/classifier training) has attracted increasing attentions due to its fundamental importance. To address this problem, many active learning (AL) methods have been proposed that employ up-to-date detectors to retrieve representative minority samples according to predefined confidence or uncertainty thresholds. However, these AL methods cause the detectors to ignore the remaining majority samples (i.e., those with low uncertainty or high prediction confidence). In this paper, by developing a principled active sample mining (ASM) framework, we demonstrate that cost-effective mining samples from these unlabeled majority data are a key to train more powerful object detectors while minimizing user effort. Specifically, our ASM framework involves a switchable sample selection mechanism for determining whether an unlabeled sample should be manually annotated via AL or automatically pseudolabeled via a novel self-learning process. The proposed process can be compatible with mini-batch-based training (i.e., using a batch of unlabeled or partially labeled data as a one-time input) for object detection. In this process, the detector, such as a deep neural network, is first applied to the unlabeled samples (i.e., object proposals) to estimate their labels and output the corresponding prediction confidences. Then, our ASM framework is used to select a number of samples and assign pseudolabels to them. These labels are specific to each learning batch based on the confidence levels and additional constraints introduced by the AL process and will be discarded afterward. Then, these temporarily labeled samples are employed for network fine-tuning. In addition, a few samples with low-confidence predictions are selected and annotated via AL. Notably, our method is suitable for object categories that are not seen in the unlabeled data during the learning process. Extensive experiments on two public benchmarks (i.e., the PASCAL VOC 2007/2012 data sets) clearly demonstrate that our ASM framework can achieve performance comparable to that of the alternative methods but with significantly fewer annotations.

Accelerating deep neural networks (DNNs) has been attracting increasing attention as it can benefit a wide range of applications, e.g., enabling mobile systems with limited computing resources to own powerful visual recognition ability. A practical strategy to this goal usually relies on a two-stage process: operating on the trained DNNs (e.g., approximating the convolutional filters with tensor decomposition) and fine-tuning the amended network, leading to difficulty in balancing the trade-off between acceleration and maintaining recognition performance. In this work, aiming at a general and comprehensive way for neural network acceleration, we develop a Wavelet-like Auto-Encoder (WAE) that decomposes the original input image into two low-resolution channels (sub-images) and incorporate the WAE into the classification neural networks for joint training. The two decomposed channels, in particular, are encoded to carry the low-frequency information (e.g., image profiles) and high-frequency (e.g., image details or noises), respectively, and enable reconstructing the original input image through the decoding process. Then, we feed the low-frequency channel into a standard classification network such as VGG or ResNet and employ a very lightweight network to fuse with the high-frequency channel to obtain the classification result. Compared to existing DNN acceleration solutions, our framework has the following advantages: i) it is tolerant to any existing convolutional neural networks for classification without amending their structures; ii) the WAE provides an interpretable way to preserve the main components of the input image for classification.

Recognizing multiple labels of images is a fundamental but challenging task in computer vision, and remarkable progress has been attained by localizing semantic-aware image regions and predicting their labels with deep convolutional neural networks. The step of hypothesis regions (region proposals) localization in these existing multi-label image recognition pipelines, however, usually takes redundant computation cost, e.g., generating hundreds of meaningless proposals with non-discriminative information and extracting their features, and the spatial contextual dependency modeling among the localized regions are often ignored or over-simplified. To resolve these issues, this paper proposes a recurrent attention reinforcement learning framework to iteratively discover a sequence of attentional and informative regions that are related to different semantic objects and further predict label scores conditioned on these regions. Besides, our method explicitly models long-term dependencies among these attentional regions that help to capture semantic label co-occurrence and thus facilitate multi-label recognition. Extensive experiments and comparisons on two large-scale benchmarks (i.e., PASCAL VOC and MS-COCO) show that our model achieves superior performance over existing state-of-the-art methods in both performance and efficiency as well as explicitly identifying image-level semantic labels to specific object regions.

Deep learning based salient object detection has recently achieved great success with its performance greatly outperforms any other unsupervised methods. However, annotating per-pixel saliency masks is a tedious and inefficient procedure. In this paper, we note that superior salient object detection can be obtained by iteratively mining and correcting the labeling ambiguity on saliency maps from traditional unsupervised methods. We propose to use the combination of a coarse salient object activation map from the classification network and saliency maps generated from unsupervised methods as pixel-level annotation, and develop a simple yet very effective algorithm to train fully convolutional networks for salient object detection supervised by these noisy annotations. Our algorithm is based on alternately exploiting a graphical model and training a fully convolutional network for model updating. The graphical model corrects the internal labeling ambiguity through spatial consistency and structure preserving while the fully convolutional network helps to correct the cross-image semantic ambiguity and simultaneously update the coarse activation map for next iteration. Experimental results demonstrate that our proposed method greatly outperforms all state-of-the-art unsupervised saliency detection methods and can be comparable to the current best strongly-supervised methods training with thousands of pixel-level saliency map annotations on all public benchmarks.

The collaborative reasoning for understanding each image-question pair is very critical but under-explored for an interpretable Visual Question Answering (VQA) system. Although very recent works also tried the explicit compositional processes to assemble multiple sub-tasks embedded in the questions, their models heavily rely on the annotations or hand-crafted rules to obtain valid reasoning layout, leading to either heavy labor or poor performance on composition reasoning. In this paper, to enable global context reasoning for better aligning image and language domains in diverse and unrestricted cases, we propose a novel reasoning network called Adversarial Composition Modular Network (ACMN). This network comprises of two collaborative modules: i) an adversarial attention module to exploit the local visual evidence for each word parsed from the question; ii) a residual composition module to compose the previously mined evidence. Given a dependency parse tree for each question, the adversarial attention module progressively discovers salient regions of one word by densely combining regions of child word nodes in an adversarial manner. Then residual composition module merges the hidden representations of an arbitrary number of children through sum pooling and residual connection. Our ACMN is thus capable of building an interpretable VQA system that gradually dives the image cues following a question-driven reasoning route and makes global reasoning by incorporating the learned knowledge of all attention modules in a principled manner. Experiments on relational datasets demonstrate the superiority of our ACMN and visualization results show the explainable capability of our reasoning system.

Image saliency detection has recently witnessed significant progress due to deep convolutional neural networks. However, extending state-of-the-art saliency detectors from image to video is challenging. The performance of salient object detection suffers from object or camera motion and the dramatic change of the appearance contrast in videos. In this paper, we present flow guided recurrent neural encoder (FGRNE), an accurate and end-to-end learning framework for video salient object detection. It works by enhancing the temporal coherence of the per-frame feature by exploiting both motion information in terms of optical flow and sequential feature evolution encoding in terms of LSTM networks. It can be considered as a universal framework to extend any FCN based static saliency detector to video salient object detection. Intensive experimental results verify the effectiveness of each part of FGRNE and confirm that our proposed method significantly outperforms state-of-the-art methods on the public benchmarks of DAVIS and FBMS.

Previous monocular depth estimation methods take a single view and directly regress the expected results. Though recent advances are made by applying geometrically inspired loss functions during training, the inference procedure does not explicitly impose any geometrical constraint. Therefore these models purely rely on the quality of data and the effectiveness of learning to generalize. This either leads to suboptimal results or the demand of huge amount of expensive ground truth labelled data to generate reasonable results. In this paper, we show for the first time that the monocular depth estimation problem can be reformulated as two sub-problems, a view synthesis procedure followed by stereo matching, with two intriguing properties, namely i) geometrical constraints can be explicitly imposed during inference; ii) demand on labelled depth data can be greatly alleviated. We show that the whole pipeline can still be trained in an end-to-end fashion and this new formulation plays a critical role in advancing the performance. The resulting model outperforms all the previous monocular depth estimation methods as well as the stereo block matching method in the challenging KITTI dataset by only using a small number of real training data. The model also generalizes well to other monocular depth estimation benchmarks. We also discuss the implications and the advantages of solving monocular depth estimation using stereo methods.

We observed that recent state-of-the-art results on single image human pose estimation were achieved by multistage Convolution Neural Networks (CNN). Notwithstanding the superior performance on static images, the application of these models on videos is not only computationally intensive, it also suffers from performance degeneration and flicking. Such suboptimal results are mainly attributed to the inability of imposing sequential geometric consistency, handling severe image quality degradation (e.g. motion blur and occlusion) as well as the inability of capturing the temporal correlation among video frames. In this paper, we proposed a novel recurrent network to tackle these problems. We showed that if we were to impose the weight sharing scheme to the multi-stage CNN, it could be re-written as a Recurrent Neural Network (RNN). This property decouples the relationship among multiple network stages and results in significantly faster speed in invoking the network for videos. It also enables the adoption of Long Short-Term Memory (LSTM) units between video frames. We found such memory augmented RNN is very effective in imposing geometric consistency among frames. It also well handles input quality degradation in videos while successfully stabilizes the sequential outputs. The experiments showed that our approach significantly outperformed current state-of-the-art methods on two large-scale video pose estimation benchmarks. We also explored the memory cells inside the LSTM and provided insights on why such mechanism would benefit the prediction for video-based pose estimations.1

Despite the recent success of stereo matching with convolutional neural networks (CNNs), it remains arduous to generalize a pre-trained deep stereo model to a novel domain. A major difficulty is to collect accurate ground-truth disparities for stereo pairs in the target domain. In this work, we propose a self-adaptation approach for CNN training, utilizing both synthetic training data (with ground-truth disparities) and stereo pairs in the new domain (without ground-truths). Our method is driven by two empirical observations. By feeding real stereo pairs of different domains to stereo models pre-trained with synthetic data, we see that: i) a pre-trained model does not generalize well to the new domain, producing artifacts at boundaries and ill-posed regions; however, ii) feeding an up-sampled stereo pair leads to a disparity map with extra details. To avoid i) while exploiting ii), we formulate an iterative optimization problem with graph Laplacian regularization. At each iteration, the CNN adapts itself better to the new domain: we let the CNN learn its own higher-resolution output; at the meanwhile, a graph Laplacian regularization is imposed to discriminatively keep the desired edges while smoothing out the artifacts. We demonstrate the effectiveness of our method in two domains: daily scenes collected by smart-phone cameras, and street views captured in a driving car.

Though quite challenging, leveraging large-scale unlabeled or partially labeled images in a cost-effective way has increasingly attracted interests for its great importance to computer vision. To tackle this problem, many Active Learning (AL) methods have been developed. However, these methods mainly define their sample selection criteria within a single image context, leading to the suboptimal robustness and impractical solution for large-scale object detection. In this paper, aiming to remedy the drawbacks of existing AL methods, we present a principled Self-supervised Sample Mining (SSM) process accounting for the real challenges in object detection. Specifically, our SSM process concentrates on automatically discovering and pseudo-labeling reliable region proposals for enhancing the object detector via the introduced cross image validation, i.e., pasting these proposals into different labeled images to comprehensively measure their values under different image contexts. By resorting to the SSM process, we propose a new AL framework for gradually incorporating unlabeled or partially labeled data into the model learning while minimizing the annotating effort of users. Extensive experiments on two public benchmarks clearly demonstrate our proposed framework can achieve the comparable performance to the state-of-the-art methods with significantly fewer annotations.

Automatically describing open-domain videos with natural language are attracting increasing interest in the field of artificial intelligence. Most existing methods simply borrow ideas from image captioning and obtain a compact video representation from an ensemble of global image feature before feeding to an RNN decoder which outputs a sentence of variable length. However, it is not only arduous for the generator to focus on specific salient objects at different time given the global video representation, it is more formidable to capture the fine-grained motion information and the relation between moving instances for more subtle linguistic descriptions. In this paper, we propose a Trajectory Structured Attentional Encoder-Decoder (TSA-ED) neural network framework for more elaborate video captioning which works by integrating local spatial-temporal representation at trajectory level through structured attention mechanism. Our proposed method is based on a LSTM-based encoder-decoder framework, which incorporates an attention modeling scheme to adaptively learn the correlation between sentence structure and the moving objects in videos, and consequently generates more accurate and meticulous statement description in the decoding stage. Experimental results demonstrate that the feature representation and structured attention mechanism based on the trajectory cluster can efficiently obtain the local motion information in the video to help generate a more fine-grained video description, and achieve the state-of-the-art performance on the well-known Charades and MSVD datasets.

Unsupervised domain adaptation (UDA) conventionally assumes labeled source samples coming from a single underlying source distribution. Whereas in practical scenario, labeled data are typically collected from diverse sources. The multiple sources are different not only from the target but also from each other, thus, domain adaptater should not be modeled in the same way. Moreover, those sources may not completely share their categories, which further brings a new transfer challenge called category shift. In this paper, we propose a deep cocktail network (DCTN) to battle the domain and category shifts among multiple sources. Motivated by the theoretical results in [33], the target distribution can be represented as the weighted combination of source distributions, and, the multi-source UDA via DCTN is then performed as two alternating steps: i) It deploys multi-way adversarial learning to minimize the discrepancy between the target and each of the multiple source domains, which also obtains the source-specific perplexity scores to denote the possibilities that a target sample belongs to different source domains. ii) The multi-source category classifiers are integrated with the perplexity scores to classify target sample, and the pseudo-labeled target samples together with source samples are utilized to update the multi-source category classifier and the feature extractor. We evaluate DCTN in three domain adaptation benchmarks, which clearly demonstrate the superiority of our framework.

We investigate a novel approach for image restoration by reinforcement learning. Unlike existing studies that mostly train a single large network for a specialized task, we prepare a toolbox consisting of small-scale convolutional networks of different complexities and specialized in different tasks. Our method, RL-Restore, then learns a policy to select appropriate tools from the toolbox to progressively restore the quality of a corrupted image. We formulate a stepwise reward function proportional to how well the image is restored at each step to learn the action policy. We also devise a joint learning scheme to train the agent and tools for better performance in handling uncertainty. In comparison to conventional human-designed networks, RL-Restore is capable of restoring images corrupted with complex and unknown distortions in a more parameter-efficient manner using the dynamically formed toolchain1.

A cognitive robot usually needs to perform multiple tasks in practice and needs to locate the desired area for each task. Since deep learning has achieved substantial progress in image recognition, to solve this area detection problem, it is straightforward to label a functional area (affordance) image dataset and apply a well-trained deep-model-based classifier on all the potential image regions. However, annotating the functional area is time consuming and the requirement of large amount of training data limits the application scope. We observe that the functional area are usually related to the surrounding object context. In this work, we propose to use the existing object detection dataset and employ the object context as effective prior to improve the performance without additional annotated data. In particular, we formulate a two-stream network that fuses the object-related and functionality-related feature for functional area detection. The whole system is formulated in an end-to-end manner and easy to implement with current object detection framework. Experiments demonstrate that the proposed network outperforms current method by almost 20% in terms of precision and recall.

Humans can naturally understand an image in depth with the aid of rich knowledge accumulated from daily lives or professions. For example, to achieve fine-grained image recognition (e.g., categorizing hundreds of subordinate categories of birds) usually requires a comprehensive visual concept organization including category labels and part-level attributes. In this work, we investigate how to unify rich professional knowledge with deep neural network architectures and propose a Knowledge-Embedded Representation Learning (KERL) framework for handling the problem of fine-grained image recognition. Specifically, we organize the rich visual concepts in the form of knowledge graph and employ a Gated Graph Neural Network to propagate node message through the graph for generating the knowledge representation. By introducing a novel gated mechanism, our KERL framework incorporates this knowledge representation into the discriminative image feature learning, i.e., implicitly associating the specific attributes with the feature maps. Compared with existing methods of fine-grained image classification, our KERL framework has several appealing properties: i) The embedded high-level knowledge enhances the feature representation, thus facilitating distinguishing the subtle differences among subordinate categories. ii) Our framework can learn feature maps with a meaningful configuration that the highlighted regions finely accord with the nodes (specific attributes) of the knowledge graph. Extensive experiments on the widely used Caltech-UCSD bird dataset demonstrate the superiority of our KERL framework over existing state-of-the-art methods.

Crowd counting from unconstrained scene images is a crucial task in many real-world applications like urban surveillance and management, but it is greatly challenged by the camera’s perspective that causes huge appearance variations in people’s scales and rotations. Conventional methods address such challenges by resorting to fixed multi-scale architectures that are often unable to cover the largely varied scales while ignoring the rotation variations. In this paper, we propose a unified neural network framework, named Deep Recurrent Spatial-Aware Network, which adaptively addresses the two issues in a learnable spatial transform module with a region-wise refinement process. Specifically, our framework incorporates a Recurrent Spatial-Aware Refinement (RSAR) module iteratively conducting two components: i) a Spatial Transformer Network that dynamically locates an attentional region from the crowd density map and transforms it to the suitable scale and rotation for optimal crowd estimation; ii) a Local Refinement Network that refines the density map of the attended region with residual learning. Extensive experiments on four challenging benchmarks show the effectiveness of our approach. Specifically, comparing with the existing best-performing methods, we achieve an improvement of 12\% on the largest dataset WorldExpo’10 and 22.8\% on the most challenging dataset UCF\_CC\_50

In this paper, we propose a two-stage depth ranking based method (DRPose3D) to tackle the problem of 3D human pose estimation. Instead of accurate 3D positions, the depth ranking can be identified by human intuitively and learned using the deep neural network more easily by solving classification problems. Moreover, depth ranking contains rich 3D information. It prevents the 2D-to-3D pose regression in two-stage methods from being ill-posed. In our method, firstly, we design a Pairwise Ranking Convolutional Neural Network (PRCNN) to extract depth rankings of human joints from images. Secondly, a coarse-to-fine 3D Pose Network(DPNet) is proposed to estimate 3D poses from both depth rankings and 2D human joint locations. Additionally, to improve the generality of our model, we introduce a statistical method to augment depth rankings. Our approach outperforms the state-of-the-art methods in the Human3.6M benchmark for all three testing protocols, indicating that depth ranking is an essential geometric feature which can be learned to improve the 3D pose estimation.

Social relationships (e.g., friends, couple etc.) form the basis of the social network in our daily life. Automatically interpreting such relationships bears a great potential for the intelligent systems to understand human behavior in depth and to better interact with people at a social level. Human beings interpret the social relationships within a group not only based on the people alone, and the interplay between such social relationships and the contextual information around the people also plays a significant role. However, these additional cues are largely overlooked by the previous studies. We found that the interplay between these two factors can be effectively modeled by a novel structured knowledge graph with proper message propagation and attention. And this structured knowledge can be efficiently integrated into the deep neural network architecture to promote social relationship understanding by an end-to-end trainable Graph Reasoning Model (GRM), in which a propagation mechanism is learned to propagate node message through the graph to explore the interaction between persons of interest and the contextual objects. Meanwhile, a graph attentional mechanism is introduced to explicitly reason about the discriminative objects to promote recognition. Extensive experiments on the public benchmarks demonstrate the superiority of our method over the existing leading competitors.

Compared to natural RGB images, data captured by 3D / depth sensors (e.g., Microsoft Kinect) have different properties, e.g., less discriminable in appearance due to lacking color / texture information. Applying convolutional neural networks (CNNs) on these depth data would lead to unsatisfying learning efficiency, i.e., requiring large amounts of annotated training data for convergence. To address this issue, this paper proposes a novel memory network module, called Convolutional Memory Block (CMB), which empowers CNNs with the memory mechanism on handling depth data. Different from the existing memory networks that store long / short term dependency from sequential data, our proposed CMB focuses on modeling the representative dependency (correlation) among non-sequential samples. Specifically, our CMB consists of one internal memory (i.e., a set of feature maps) and three specific controllers, which enable a powerful yet efficient memory manipulation mechanism. In this way, the internal memory, being implicitly aggregated from all previous inputted samples, can learn to store and utilize representative features among the samples. Furthermore, we employ our CMB to develop a concise framework for predicting articulated pose from still depth images. Comprehensive evaluations on three public benchmarks demonstrate significant superiority (about 6%) of our framework over all the compared methods. More importantly, thanks to the enhanced learning efficiency, our framework can still achieve satisfying results using 50% less training data.

Dense mapping is always the desire of simultaneous localization and mapping (SLAM), especially for the applications that require fast and dense scene information. Visual-inertial odometry (VIO) is a light-weight and effective solution to fast self-localization. However, VIO-based SLAM systems have difficulty in providing dense mapping results due to the spatial sparsity and temporal instability of the VIO depth estimations. Although there have been great efforts on real-time mapping and depth recovery from sparse measurements, the existing solutions for VIO-based SLAM still fail to preserve sufficient geometry details in their results. In this paper, we propose to embed depth recovery into VIO-based SLAM for real-time dense mapping. In the proposed method, we present a subspace-based stabilization scheme to maintain the temporal consistency and design a hierarchical pipeline for edge-preserving depth interpolation to reduce the computational burden. Numerous experiments demonstrate that our method can achieve an accuracy improvement of up to 49.1 cm compared to state-of-the-art learning-based methods for depth recovery and reconstruct sufficient geometric details in dense mapping when only 0.07% depth samples are available. Since a simple CPU implementation of our method already runs at 10-20 fps, we believe our method is very favorable for practical SLAM systems with critical computational requirements.

Object categories inherently form a hierarchy with different levels of concept abstraction, especially for fine-grained categories. For example, birds (Aves) can be categorized according to a four-level hierarchy of order, family, genus, and species. This hierarchy encodes rich correlations among various categories across different levels, which can effectively regularize the semantic space and thus make prediction less ambiguous. However, previous studies of fine-grained image recognition primarily focus on categories of one certain level and usually overlook this correlation information. In this work, we investigate simultaneously predicting categories of different levels in the hierarchy and integrating this structured correlation information into the deep neural network by developing a novel Hierarchical Semantic Embedding (HSE) framework. Specifically, the HSE framework sequentially predicts the category score vector of each level in the hierarchy, from highest to lowest. At each level, it incorporates the predicted score vector of the higher level as prior knowledge to learn finer-grained feature representation. During training, the predicted score vector of the higher level is also employed to regularize label prediction by using it as soft targets of corresponding sub-categories. To evaluate the proposed framework, we organize the 200 bird species of the Caltech-UCSD birds dataset with the four-level category hierarchy and construct a large-scale butterfly dataset that also covers four level categories. Extensive experiments on these two and the newly-released VegFru datasets demonstrate the superiority of our HSE framework over the baseline methods and existing competitors.

Single image rain streaks removal has recently witnessed substantial progress due to the development of deep convolutional neural networks. However, existing deep learning based methods either focus on the entrance and exit of the network by decomposing the input image into high and low frequency information and employing residual learning to reduce the mapping range, or focus on the introduction of cascaded learning scheme to decompose the task of rain streaks removal into multi-stages. These methods treat the convolutional neural network as an encapsulated end-to-end mapping module without deepening into the rationality and superiority of neural network design. In this paper, we delve into an effective end-to-end neural network structure for stronger feature expression and spatial correlation learning. Specifically, we propose a non-locally enhanced encoder-decoder network framework, which consists of a pooling indices embedded encoder-decoder network to efficiently learn increasingly abstract feature representation for more accurate rain streaks modeling while perfectly preserving the image detail. The proposed encoder-decoder framework is composed of a series of non-locally enhanced dense blocks that are designed to not only fully exploit hierarchical features from all the convolutional layers but also well capture the long-distance dependencies and structural information. Extensive experiments on synthetic and real datasets demonstrate that the proposed method can effectively remove rain-streaks on rainy image of various densities while well preserving the image details, which achieves significant improvements over the recent state-of-the-art methods.

Facial makeup transfer aims to translate the makeup style from a given reference makeup face image to another non-makeup one while preserving face identity. Such an instance-level transfer problem is more challenging than conventional domain-level transfer tasks, especially when paired data is unavailable. Makeup style is also different from global styles (e.g., paintings) in that it consists of several local styles/cosmetics, including eye shadow, lipstick, foundation, and so on. Extracting and transferring such local and delicate makeup information is infeasible for existing style transfer methods. We address the issue by incorporating both global domain-level loss and local instance-level loss in an dual input/output Generative Adversarial Network, called BeautyGAN. Specifically, the domain-level transfer is ensured by discriminators that distinguish generated images from domains' real samples. The instance-level loss is calculated by pixel-level histogram loss on separate local facial regions. We further introduce perceptual loss and cycle consistency loss to generate high quality faces and preserve identity. The overall objective function enables the network to learn translation on instance-level through unsupervised adversarial learning. We also build up a new makeup dataset that consists of 3834 high-resolution face images. Extensive experiments show that BeautyGAN could generate visually pleasant makeup faces and accurate transferring results. Data and code are available at http://liusi-group.com/projects/BeautyGAN.

Traffic flow prediction is crucial for urban traffic management and public safety. Its key challenges lie in how to adaptively integrate the various factors that affect the flow changes. In this paper, we propose a unified neural network module to address this problem, called Attentive Crowd Flow Machine~(ACFM), which is able to infer the evolution of the crowd flow by learning dynamic representations of temporally-varying data with an attention mechanism. Specifically, the ACFM is composed of two progressive ConvLSTM units connected with a convolutional layer for spatial weight prediction. The first LSTM takes the sequential flow density representation as input and generates a hidden state at each time-step for attention map inference, while the second LSTM aims at learning the effective spatial-temporal feature expression from attentionally weighted crowd flow features. Based on the ACFM, we further build a deep architecture with the application to citywide crowd flow prediction, which naturally incorporates the sequential and periodic data as well as other external influences. Extensive experiments on two standard benchmarks (i.e., crowd flow in Beijing and New York City) show that the proposed method achieves significant improvements over the state-of-the-art methods.

This article summarizes the corresponding half-day tutorial at ACM Multimedia 2018. This tutorial reviews recent progresses for pixel-level understanding with structured deep learning, including 1) human-centric analysis: human parsing and pose estimation; 2) part-based analysis: object part and face parsing; 3) weakly-supervised analysis: object localization and semantic segmentation; 4) depth estimation: stereo matching.

Abstract The dominant object detection approaches treat the recognition of each region separately and overlook crucial semantic correlations between objects in one scene. This paradigm leads to substantial performance drop when facing heavy long-tail problems, where very few samples are available for rare classes and plenty of confusing categories exists. We exploit diverse human commonsense knowledge for reasoning over large-scale object categories and reaching semantic coherency within one image. Particularly, we present Hybrid Knowledge Routed Modules (HKRM) that incorporates the reasoning routed by two kinds of knowledge forms: an explicit knowledge module for structured constraints that are summarized with linguistic knowledge (e.g. shared attributes, relationships) about concepts; and an implicit knowledge module that depicts some implicit constraints (e.g. common spatial layouts). By functioning over a region-to-region graph, both modules can be individualized and adapted to coordinate with visual patterns in each image, guided by specific knowledge forms. HKRM are light-weight, general-purpose and extensible by easily incorporating multiple knowledge to endow any detection networks the ability of global semantic reasoning. Experiments on large-scale object detection benchmarks show HKRM obtains around 34.5% improvement on VisualGenome (1000 categories) and 30.4% on ADE in terms of mAP.

Beyond local convolution networks, we explore how to harness various external human knowledge for endowing the networks with the capability of semantic global reasoning. Rather than using separate graphical models (e.g. CRF) or constraints for modeling broader dependencies, we propose a new Symbolic Graph Reasoning (SGR) layer, which performs reasoning over a group of symbolic nodes whose outputs explicitly represent different properties of each semantic in a prior knowledge graph. To cooperate with local convolutions, each SGR is constituted by three modules: a) a primal local-to-semantic voting module where the features of all symbolic nodes are generated by voting from local representations; b) a graph reasoning module propagates information over knowledge graph to achieve global semantic coherency; c) a dual semantic-to-local mapping module learns new associations of the evolved symbolic nodes with local representations, and accordingly enhances local features. The SGR layer can be injected between any convolution layers and instantiated with distinct prior graphs. Extensive experiments show incorporating SGR significantly improves plain ConvNets on three semantic segmentation tasks and one image classification task. More analyses show the SGR layer learns shared symbolic representations for domains/datasets with the different label set given a universal knowledge graph, demonstrating its superior generalization capability.

As an indispensable component, Batch Normalization (BN) has successfully improved the training of deep neural networks (DNNs) with mini-batches, by normalizing the distribution of the internal representation for each hidden layer. However, the effectiveness of BN would diminish with the scenario of micro-batch (e.g. less than 4 samples in a mini-batch), since the estimated statistics in a mini-batch are not reliable with insufficient samples. This limits BN's room in training larger models on segmentation, detection, and video-related problems, which require small batches constrained by memory consumption. In this paper, we present a novel normalization method, called Kalman Normalization (KN), for improving and accelerating the training of DNNs, particularly under the context of micro-batches. Specifically, unlike the existing solutions treating each hidden layer as an isolated system, KN treats all the layers in a network as a whole system, and estimates the statistics of a certain layer by considering the distributions of all its preceding layers, mimicking the merits of Kalman Filtering. On ResNet50 trained in ImageNet, KN has 3.4% lower error than its BN counterpart when using a batch size of 4; Even when using typical batch sizes, KN still maintains an advantage over BN while other BN variants suffer a performance degradation. Moreover, KN can be naturally generalized to many existing normalization variants to obtain gains, e.g. equipping Group Normalization with Group Kalman Normalization (GKN). KN can outperform BN and its variants for large scale object detection and segmentation task in COCO 2017.

Instance-level object segmentation is an important yet under-explored task. Most of state-of-the-art methods rely on region proposal methods to extract candidate segments and then utilize object classification to produce final results. Nonetheless, generating reliable region proposals itself is a quite challenging and unsolved task. In this work, we propose a Proposal-Free Network (PFN) to address the instance-level object segmentation problem, which outputs the numbers of instances of different categories and the pixel-level information on i) the coordinates of the instance bounding box each pixel belongs to, and ii) the confidences of different categories for each pixel, based on pixel-to-pixel deep convolutional neural network. All the outputs together, by using any off-the-shelf clustering method for simple post-processing, can naturally generate the ultimate instance-level object segmentation results. The whole PFN can be easily trained without the requirement of a proposal generation stage. Extensive evaluations on the challenging PASCAL VOC 2012 semantic segmentation benchmark demonstrate the effectiveness of the proposed PFN solution without relying on any proposal generation methods.

This paper aims to develop a novel cost-effective framework for face identification, which progressively maintains a batch of classifiers with the increasing face images of different individuals. By naturally combining two recently rising techniques: active learning (AL) and self-paced learning (SPL), our framework is capable of automatically annotating new instances and incorporating them into training under weak expert recertification. We first initialize the classifier using a few annotated samples for each individual, and extract image features using the convolutional neural nets. Then, a number of candidates are selected from the unannotated samples for classifier updating, in which we apply the current classifiers ranking the samples by the prediction confidence. In particular, our approach utilizes the high-confidence and low-confidence samples in the self-paced and the active user-query way, respectively. The neural nets are later fine-tuned based on the updated classifiers. Such heuristic implementation is formulated as solving a concise active SPL optimization problem, which also advances the SPL development by supplementing a rational dynamic curriculum constraint. The new model finely accords with the "instructor-student-collaborative" learning mode in human education. The advantages of this proposed framework are two-folds: i) The required number of annotated samples is significantly decreased while the comparable performance is guaranteed. A dramatic reduction of user effort is also achieved over other state-of-the-art active learning techniques. ii) The mixture of SPL and AL effectively improves not only the classifier accuracy compared to existing AL/SPL methods but also the robustness against noisy data. We evaluate our framework on two challenging datasets, which include hundreds of persons under diverse conditions, and demonstrate very promising results. Please find the code of this project at: http://hcp.sysu.edu.cn/projects/aspl/.

Aiming at improving the performance of visual classification in a cost-effective manner, this paper proposes an incremental semi-supervised learning paradigm called deep co-space (DCS). Unlike many conventional semi-supervised learning methods usually performed within a fixed feature space, our DCS gradually propagates information from labeled samples to unlabeled ones along with deep feature learning. We regard deep feature learning as a series of steps pursuing feature transformation, i.e., projecting the samples from a previous space into a new one, which tends to select the reliable unlabeled samples with respect to this setting. Specifically, for each unlabeled image instance, we measure its reliability by calculating the category variations of feature transformation from two different neighborhood variation perspectives and merged them into a unified sample mining criterion deriving from Hellinger distance. Then, those samples keeping stable correlation to their neighboring samples (i.e., having small category variation in distribution) across the successive feature space transformation are automatically received labels and incorporated into the model for incrementally training in terms of classification. Our extensive experiments on standard image classification benchmarks (e.g., Caltech-256 and SUN-397) demonstrate that the proposed framework is capable of effectively mining from large-scale unlabeled images, which boosts image classification performance and achieves promising results compared with other semi-supervised learning methods.

Learning similarity and distance measures has become increasingly important for the analysis, matching, retrieval, recognition, and categorization of video and multimedia data. With the ubiquitous use of digital imaging devices, mobile terminals and social networks, there are massive volumes of heterogeneous and homogeneous video and multimedia data from multiple sources, views, and domains, e.g., news media websites, microblog, mobile phone, social networking, etc. Similarity and distance-based constraints can also be extended and incorporated to boost classification and relationship learning. Moreover, the spatio-temporal coherence among video data can also be utilized for self-supervised learning of similarity and distance metrics. This trend has brought several challenging issues for developing similarity and metric learning methods for large scale and weakly annotated data, where outliers and incorrectly annotated data are inevitable. Recently, scalability has been investigated to cope with lightweight and large scale metric learning, while nonlinear similarity models have shown their great potentials in learning invariant representation and nonlinear measures of video and multimedia data.

To avoid the exhaustive search over locations and scales, current state-of-the-art object detection systems usually involve a crucial component generating a batch of candidate object proposals from images. In this paper, we present a simple yet effective approach for segmenting object proposals via a deep architecture of recursive neural networks (ReNNs), which hierarchically groups regions for detecting object candidates over scales. Unlike traditional methods that mainly adopt fixed similarity measures for merging regions or finding object proposals, our approach adaptively learns the region merging similarity and the objectness measure during the process of hierarchical region grouping. Specifically, guided by a structured loss, the ReNN model jointly optimizes the cross-region similarity metric with the region merging process as well as the objectness prediction. During inference of the object proposal generation, we introduce randomness into the greedy search to cope with the ambiguity of grouping regions. Extensive experiments on standard benchmarks, e.g., PASCAL VOC and ImageNet, suggest that our approach is capable of producing object proposals with high recall while well preserving the object boundaries and outperforms other existing methods in both accuracy and efficiency.

This article presents a high-precision multi-modal approach for localizing moving cameras with monocular videos, which has wide potentials in many intelligent applications, including robotics, autonomous vehicles, and so on. Existing visual odometry methods often suffer from symmetric or repetitive scene patterns, e.g., windows on buildings or parking stalls. To address this issue, we introduce a robust camera localization method that contributes in two aspects. First, we formulate feature tracking, the critical step of visual odometry, as a hierarchical min-cost network flow optimization task, and we regularize the formula with flow constraints, cross-scale consistencies, and motion heuristics. The proposed regularized formula is capable of adaptively selecting distinctive features or feature combinations, which is more effective than traditional methods that detect and group repetitive patterns in a separate step. Second, we develop a joint formula for integrating dense visual odometry and sparse GPS readings in a common reference coordinate. The fusion process is guided with high-order statistics knowledge to suppress the impacts of noises, clusters, and model drifting. We evaluate the proposed camera localization method on both public video datasets and a newly created dataset that includes scenes full of repetitive patterns. Results with comparisons show that our method can achieve comparable performance to state-of-the-art methods and is particularly effective for addressing repetitive pattern issues.

The generalization error bound of the support vector machine (SVM) depends on the ratio of the radius and margin. However, conventional SVM only considers the maximization of the margin but ignores the minimization of the radius, which restricts its performance when applied to joint learning of feature transformation and the SVM classifier. Although several approaches have been proposed to integrate the radius and margin information, most of them either require the form of the transformation matrix to be diagonal, or are nonconvex and computationally expensive. In this paper, we suggest a novel approximation for the radius of the minimum enclosing ball in feature space, and then propose a convex radius-margin-based SVM model for joint learning of feature transformation and the SVM classifier, i.e., F-SVM. A generalized block coordinate descent method is adopted to solve the F-SVM model, where the feature transformation is updated via the gradient descent and the classifier is updated by employing the existing SVM solver. By incorporating with kernel principal component analysis, F-SVM is further extended for joint learning of nonlinear transformation and the classifier. F-SVM can also be incorporated with deep convolutional networks to improve image classification performance. Experiments on the UCI, LFW, MNIST, CIFAR-10, CIFAR-100, and Caltech101 data sets demonstrate the effectiveness of F-SVM.

Existing visual tracking methods usually localize the object with a bounding box, in which the foreground object trackers/detectors are often disturbed by the introduced background information. To handle this problem, we aim to learn a more robust object representation for visual tracking. In particular, the tracked object is represented with a graph structure (i.e., a set of non-overlapping image patches), in which the weight of each node (patch) indicates how likely it belongs to the foreground and edges are also weighed for indicating the appearance compatibility of two neighboring nodes. This graph is dynamically learnt (i.e., the nodes and edges received weights) and applied in object tracking and model updating. We constrain the graph learning from two aspects: i) the global low-rank structure over all nodes and ii) the local sparseness of node neighbors. During the tracking process, our method performs the following steps at each frame. First, the graph is initialized by assigning either 1 or 0 to the weights of some image patches according to the predicted bounding box. Second, the graph is optimized through designing a new ALM (Augmented Lagrange Multiplier) based algorithm. Third, the object feature representation is updated by imposing the weights of patches on the extracted image features. The object location is finally predicted by adopting the Struck tracker. Extensive experiments show that our approach outperforms the state-of-the-art tracking methods on two standard benchmarks, i.e., OTB100 and NUS-PRO.

Face hallucination is a domain-specific super-resolution problem with the goal to generate high-resolution (HR) faces from low-resolution (LR) input images. In contrast to existing methods that often learn a single patch-to-patch mapping from LR to HR images and are regardless of the contextual interdependency between patches, we propose a novel Attention-aware Face Hallucination (Attention-FH) framework which resorts to deep reinforcement learning for sequentially discovering attended patches and then performing the facial part enhancement by fully exploiting the global interdependency of the image. Specifically, in each time step, the recurrent policy network is proposed to dynamically specify a new attended region by incorporating what happened in the past. The state (i.e., face hallucination result for the whole image) can thus be exploited and updated by the local enhancement network on the selected region. The Attention-FH approach jointly learns the recurrent policy network and local enhancement network through maximizing the long-term reward that reflects the hallucination performance over the whole image. Therefore, our proposed Attention-FH is capable of adaptively personalizing an optimal searching path for each face image according to its own characteristic. Extensive experiments show our approach significantly surpasses the state-of-the-arts on in-the-wild faces with large pose and illumination variations. The state (i.e., face hallucination result for the whole image) can thus be exploited and updated by the local enhancement network on the selected region. The Attention-FH approach jointly learns the recurrent policy network and local enhancement network through maximizing the long-term reward that reflects the hallucination performance over the whole image. Therefore, our proposed Attention-FH is capable of adaptively personalizing an optimal searching path for each face image according to its own characteristic. Extensive experiments show our approach significantly surpasses the state-of-the-arts on in-the-wild faces with large pose and illumination variations.

Human parsing has recently attracted a lot of research interests due to its huge application potentials. However existing datasets have limited number of images and annotations, and lack the variety of human appearances and the coverage of challenging cases in unconstrained environment. In this paper, we introduce a new benchmark Look into Person (LIP) that makes a significant advance in terms of scalability, diversity and difficulty, a contribution that we feel is crucial for future developments in human-centric analysis. This comprehensive dataset contains over 50,000 elaborately annotated images with 19 semantic part labels, which are captured from a wider range of viewpoints, occlusions and background complexity. Given these rich annotations we perform detailed analysis of the leading human parsing approaches, gaining insights into the success and failures of these methods. Furthermore, in contrast to the existing efforts on improving the feature discriminative capability, we solve human parsing by exploring a novel self-supervised structure-sensitive learning approach, which imposes human pose structures into parsing results without resorting to extra supervision (i.e., no need for specifically labeling human joints in model training). Our self-supervised learning framework can be injected into any advanced neural networks to help incorporate rich high-level knowledge regarding human joints from a global perspective and improve the parsing results. Extensive evaluations on our LIP and the public PASCAL-Person-Part dataset demonstrate the superiority of our method.

Image saliency detection has recently witnessed rapid progress due to deep convolutional neural networks. However, none of the existing methods is able to identify object instances in the detected salient regions. In this paper, we present a salient instance segmentation method that produces a saliency mask with distinct object instance labels for an input image. Our method consists of three steps, estimating saliency map, detecting salient object contours and identifying salient object instances. For the first two steps, we propose a multiscale saliency refinement network, which generates high-quality salient region masks and salient object contours. Once integrated with multiscale combinatorial grouping and a MAP-based subset optimization framework, our method can generate very promising salient object instance segmentation results. To promote further research and evaluation of salient instance segmentation, we also construct a new database of 1000 images and their pixelwise salient instance annotations. Experimental results demonstrate that our proposed method is capable of achieving state-of-the-art performance on all public benchmarks for salient region detection as well as on our new dataset for salient instance segmentation.

This paper develops a general framework for learning interpretable data representation via Long Short-Term Memory (LSTM) recurrent neural networks over hierarchal graph structures. Instead of learning LSTM models over the pre-fixed structures, we propose to further learn the intermediate interpretable multi-level graph structures in a progressive and stochastic way from data during the LSTM network optimization. We thus call this model the structure-evolving LSTM. In particular, starting with an initial element-level graph representation where each node is a small data element, the structure-evolving LSTM gradually evolves the multi-level graph representations by stochastically merging the graph nodes with high compatibilities along the stacked LSTM layers. In each LSTM layer, we estimate the compatibility of two connected nodes from their corresponding LSTM gate outputs, which is used to generate a merging probability. The candidate graph structures are accordingly generated where the nodes are grouped into cliques with their merging probabilities. We then produce the new graph structure with a Metropolis-Hasting algorithm, which alleviates the risk of getting stuck in local optimums by stochastic sampling with an acceptance probability. Once a graph structure is accepted, a higher-level graph is then constructed by taking the partitioned cliques as its nodes. During the evolving process, representation becomes more abstracted in higher-levels where redundant information is filtered out, allowing more efficient propagation of long-range data dependencies. We evaluate the effectiveness of structure-evolving LSTM in the application of semantic object parsing and demonstrate its advantage over state-of-the-art LSTM models on standard benchmarks.

3D Human articulated pose recovery from monocular image sequences is very challenging due to the diverse appearances, viewpoints, occlusions, and also the human 3D pose is inherently ambiguous from the monocular imagery. It is thus critical to exploit rich spatial and temporal long-range dependencies among body joints for accurate 3D pose sequence prediction. Existing approaches usually manually design some elaborate prior terms and human body kinematic constraints for capturing structures, which are often insufficient to exploit all intrinsic structures and not scalable for all scenarios. In contrast, this paper presents a Recurrent 3D Pose Sequence Machine(RPSM) to automatically learn the image-dependent structural constraint and sequence-dependent temporal context by using a multi-stage sequential refinement. At each stage, our RPSM is composed of three modules to predict the 3D pose sequences based on the previously learned 2D pose representations and 3D poses: (i) a 2D pose module extracting the image-dependent pose representations, (ii) a 3D pose recurrent module regressing 3D poses and (iii) a feature adaption module serving as a bridge between module (i) and (ii) to enable the representation transformation from 2D to 3D domain. These three modules are then assembled into a sequential prediction framework to refine the predicted poses with multiple recurrent stages. Extensive evaluations on the Human3.6M dataset and HumanEva-I dataset show that our RPSM outperforms all state-of-the-art approaches for 3D pose estimation.

Recent advanced deep convolutional networks (CNNs) achieved great successes in many computer vision tasks, because of their compelling learning complexity and the presences of large-scale labeled data. However, as obtaining per-pixel annotations is expensive, performances of CNNs in semantic image segmentation are not fully exploited. This work significantly increases segmentation accuracy of CNNs by learning from an Image Descriptions in the Wild (IDW) dataset. Unlike previous image captioning datasets, where captions were manually and densely annotated, images and their descriptions in IDW are automatically downloaded from Internet without any manual cleaning and refinement. An IDW-CNN is proposed to jointly train IDW and existing image segmentation dataset such as Pascal VOC 2012 (VOC). It has two appealing properties. First, knowledge from different datasets can be fully explored and transferred from each other to improve performance. Second, segmentation accuracy in VOC can be constantly increased when selecting more data from IDW. Extensive experiments demonstrate the effectiveness and scalability of IDW-CNN, which outperforms existing best-performing system by 12% on VOC12 test set.

Existing person re-identification benchmarks and methods mainly focus on matching cropped pedestrian images between queries and candidates. However, it is different from real-world scenarios where the annotations of pedestrian bounding boxes are unavailable and the target person needs to be searched from a gallery of whole scene images. To close the gap, we propose a new deep learning framework for person search. Instead of breaking it down into two separate tasks-pedestrian detection and person re-identification, we jointly handle both aspects in a single convolutional neural network. An Online Instance Matching (OIM) loss function is proposed to train the network effectively, which is scalable to datasets with numerous identities. To validate our approach, we collect and annotate a large-scale benchmark dataset for person search. It contains 18,184 images, 8,432 identities, and 96,143 pedestrian bounding boxes. Experiments show that our framework outperforms other separate approaches, and the proposed OIM loss function converges much faster and better than the conventional Softmax loss.

Deep neural networks have advanced many computer vision tasks, because of their compelling capacities to learn from large amount of labeled data. However, their performances are not fully exploited in semantic image segmentation as the scale of training set is limited, where per-pixel labelmaps are expensive to obtain. To reduce labeling efforts, a natural solution is to collect additional images from Internet that are associated with image-level tags. Unlike existing works that treated labelmaps and tags as independent supervisions, we present a novel learning setting, namely dual image segmentation (DIS), which consists of two complementary learning problems that are jointly solved. One predicts labelmaps and tags from images, and the other reconstructs the images using the predicted labelmaps. DIS has three appealing properties. 1) Given an image with tags only, its labelmap can be inferred by leveraging the images and tags as constraints. The estimated labelmaps that capture accurate object classes and boundaries are used as ground truths in training to boost performance. 2) DIS is able to clean tags that have noises. 3) DIS significantly reduces the number of per-pixel annotations in training, while still achieves state-of-the-art performance. Extensive experiments demonstrate the effectiveness of DIS, which outperforms an existing best-performing baseline by 12.6% on Pascal VOC 2012 test set, without any post-processing such as CRF/MRF smoothing.

This paper proposes a novel deep architecture to address multi-label image recognition, a fundamental and practical task towards general visual understanding. Current solutions for this task usually rely on an extra step of extracting hypothesis regions (i.e., region proposals), resulting in redundant computation and sub-optimal performance. In this work, we achieve the interpretable and contextualized multi-label image classification by developing a recurrent memorized-attention module. This module consists of two alternately performed components: i) a spatial transformer layer to locate attentional regions from the convolutional feature maps in a region-proposal-free way and ii) an LSTM (Long-Short Term Memory) sub-network to sequentially predict semantic labeling scores on the located regions while capturing the global dependencies of these regions. The LSTM also output the parameters for computing the spatial transformer. On large-scale benchmarks of multi-label image classification (e.g., MS-COCO and PASCAL VOC 07), our approach demonstrates superior performances over other existing state-of-the-arts in both accuracy and efficiency.

This paper aims at task-oriented action prediction, i.e., predicting a sequence of actions towards accomplishing a specific task under a certain scene, which is a new problem in computer vision research. The main challenges lie in how to model task-specific knowledge and integrate it in the learning procedure. In this work, we propose to train a recurrent long-short term memory (LSTM) network for handling this problem, i.e., taking a scene image (including pre-located objects) and the specified task as input and recurrently predicting action sequences. However, training such a network usually requires large amounts of annotated samples for covering the semantic space (e.g., diverse action decomposition and ordering). To alleviate this issue, we introduce a temporal And-Or graph (AOG) for task description, which hierarchically represents a task into atomic actions. With this AOG representation, we can produce many valid samples (i.e., action sequences according with common sense) by training another auxiliary LSTM network with a small set of annotated samples. And these generated samples (i.e., task-oriented action sequences) effectively facilitate training the model for task-oriented action prediction. In the experiments, we create a new dataset containing diverse daily tasks and extensively evaluate the effectiveness of our approach.

This paper presents a unified framework to learn to quantify perceptual attributes (e.g., safety, attractiveness) of physical urban environments using crowd-sourced street-view photos without human annotations. The efforts of this work include two folds. First, we collect a large-scale urban image dataset in multiple major cities in U.S.A., which consists of multiple street-view photos for every place. Instead of using subjective annotations as in previous works, which are neither accurate nor consistent, we collect for every place the safety score from government's crime event records as objective safety indicators. Second, we observe that the place-centric perception task is by nature a multi-instance regression problem since the labels are only available for places (bags), rather than images or image regions (instances). We thus introduce a deep convolutional neural network (CNN) to parameterize the instance-level scoring function, and develop an EM algorithm to alternatively estimate the primary instances (images or image regions) which affect the safety scores and train the proposed network. Our method is capable of localizing interesting images and image regions for each place. We evaluate the proposed method on a newly created dataset and a public dataset. Results with comparisons showed that our method can clearly outperform the alternative perception methods and more importantly, is capable of generating region-level safety scores to facilitate interpretations of the perception process.

An intuition on human segmentation is that when a human is moving in a video, the video-context (e.g., appearance and motion clues) may potentially infer reasonable mask information for the whole human body. Inspired by this, based on popular deep convolutional neural networks (CNN), we explore a very-weakly supervised learning framework for human segmentation task, where only an imperfect human detector is available along with massive weakly-labeled YouTube videos. In our solution, the video-context guided human mask inference and CNN based segmentation network learning iterate to mutually enhance each other until no further improvement gains. In the first step, each video is decomposed into supervoxels by the unsupervised video segmentation. The superpixels within the supervoxels are then classified as human or non-human by graph optimization with unary energies from the imperfect human detection results and the predicted confidence maps by the CNN trained in the previous iteration. In the second step, the video-context derived human masks are used as direct labels to train CNN. Extensive experiments on the challenging PASCAL VOC 2012 semantic segmentation benchmark demonstrate that the proposed framework has already achieved superior results than all previous weakly-supervised methods with object class or bounding box annotations. In addition, by augmenting with the annotated masks from PASCAL VOC 2012, our method reaches a new state-of-the-art performance on the human segmentation task.

In this work, we address the human parsing task with a novel Contextualized Convolutional Neural Network (Co-CNN) architecture, which well integrates the cross-layer context, global image-level context, semantic edge context, within-super-pixel context and cross-super-pixel neighborhood context into a unified network. Given an input human image, Co-CNN produces the pixelwise categorization in an end-to-end way. First, the cross-layer context is captured by our basic local-to-global-to-local structure, which hierarchically combines the global semantic information and the local fine details across different convolutional layers. Second, the global image-level label prediction is used as an auxiliary objective in the intermediate layer of the Co-CNN, and its outputs are further used for guiding the feature learning in subsequent convolutional layers to leverage the global image-level context. Third, semantic edge context is further incorporated into Co-CNN, where the high-level semantic boundaries are leveraged to guide pixel-wise labeling. Finally, to further utilize the local super-pixel contexts, the within-super-pixel smoothing and cross-super-pixel neighbourhood voting are formulated as natural sub-components of the Co-CNN to achieve the local label consistency in both training and testing process. Comprehensive evaluations on two public datasets well demonstrate the significant superiority of our Co-CNN over other state-of-the-arts for human parsing. In particular, the F-1 score on the large dataset [1] reaches 81.72 percent by Co-CNN, significantly higher than 62.81 percent and 64.38 percent by the state-of-the-art algorithms, M-CNN [2] and ATR [1], respectively. By utilizing our newly collected large dataset for training, our Co-CNN can achieve 85.36 percent in F-1 score.

Cross-domain visual data matching is one of the fundamental problems in many real-world vision tasks, e.g., matching persons across ID photos and surveillance videos. Conventional approaches to this problem usually involves two steps: i) projecting samples from different domains into a common space, and ii) computing (dis-)similarity in this space based on a certain distance. In this paper, we present a novel pairwise similarity measure that advances existing models by i) expanding traditional linear projections into affine transformations and ii) fusing affine Mahalanobis distance and Cosine similarity by a data-driven combination. Moreover, we unify our similarity measure with feature representation learning via deep convolutional neural networks. Specifically, we incorporate the similarity measure matrix into the deep architecture, enabling an end-to-end way of model optimization. We extensively evaluate our generalized similarity model in several challenging cross-domain matching tasks: person re-identification under different views and face verification over different modalities (i.e., faces from still images and videos, older and younger faces, and sketch and photo portraits). The experimental results demonstrate superior performance of our model over other state-of-the-art methods.

This paper investigates how to fuse grayscale and thermal video data for detecting foreground objects in challenging scenarios. To this end, we propose an intuitive yet effective method called weighted low-rank decomposition (WELD), which adaptively pursues the cross-modality low-rank representation. Specifically, we form two data matrices by accumulating sequential frames from the grayscale and the thermal videos, respectively. Within these two observing matrices, WELD detects moving foreground pixels as sparse outliers against the low-rank structure background and incorporates the weight variables to make the models of two modalities complementary to each other. The smoothness constraints of object motion are also introduced in WELD to further improve the robustness to noises. For optimization, we propose an iterative algorithm to efficiently solve the low-rank models with three subproblems. Moreover, we utilize an edge-preserving filtering-based method to substantially speed up WELD while preserving its accuracy. To provide a comprehensive evaluation benchmark of grayscale-thermal foreground detection, we create a new data set including 25 aligned grayscale-thermal video pairs with high diversity. Our extensive experiments on both the newly created data set and the public data set OSU3 suggest that WELD achieves superior performance and comparable efficiency against other state-of-the-art approaches.

Recent successes in learning-based image classification, however, heavily rely on the large number of annotated training samples, which may require considerable human effort. In this paper, we propose a novel active learning (AL) framework, which is capable of building a competitive classifier with optimal feature representation via a limited amount of labeled training instances in an incremental learning manner. Our approach advances the existing AL methods in two aspects. First, we incorporate deep convolutional neural networks into AL. Through the properly designed framework, the feature representation and the classifier can be simultaneously updated with progressively annotated informative samples. Second, we present a cost-effective sample selection strategy to improve the classification performance with less manual annotations. Unlike traditional methods focusing on only the uncertain samples of low prediction confidence, we especially discover the large amount of high-confidence samples from the unlabeled set for feature learning. Specifically, these high-confidence samples are automatically selected and iteratively assigned pseudolabels. We thus call our framework cost-effective AL (CEAL) standing for the two advantages. Extensive experiments demonstrate that the proposed CEAL framework can achieve promising results on two challenging image classification data sets, i.e., face recognition on the cross-age celebrity face recognition data set database and object categorization on Caltech-256.

Sketch portrait generation benefits a wide range of applications such as digital entertainment and law enforcement. Although plenty of efforts have been dedicated to this task, several issues still remain unsolved for generating vivid and detail-preserving personal sketch portraits. For example, quite a few artifacts may exist in synthesizing hairpins and glasses, and textural details may be lost in the regions of hair or mustache. Moreover, the generalization ability of current systems is somewhat limited since they usually require elaborately collecting a dictionary of examples or carefully tuning features/components. In this paper, we present a novel representation learning framework that generates an end-to-end photo-sketch mapping through structure and texture decomposition. In the training stage, we first decompose the input face photo into different components according to their representational contents (i.e., structural and textural parts) by using a pre-trained convolutional neural network (CNN). Then, we utilize a branched fully CNN for learning structural and textural representations, respectively. In addition, we design a sorted matching mean square error metric to measure texture patterns in the loss function. In the stage of sketch rendering, our approach automatically generates structural and textural representations for the input photo and produces the final result via a probabilistic fusion scheme. Extensive experiments on several challenging benchmarks suggest that our approach outperforms example-based synthesis algorithms in terms of both perceptual and objective metrics. In addition, the proposed method also has better generalization ability across data set without additional training.

Distance metric learning aims to learn from the given training data a valid distance metric, with which the similarity between data samples can be more effectively evaluated for classification. Metric learning is often formulated as a convex or nonconvex optimization problem, while most existing methods are based on customized optimizers and become inefficient for large scale problems. In this paper, we formulate metric learning as a kernel classification problem with the positive semi-definite constraint, and solve it by iterated training of support vector machines (SVMs). The new formulation is easy to implement and efficient in training with the off-the-shelf SVM solvers. Two novel metric learning models, namely positive-semidefinite constrained metric learning (PCML) and nonnegative-coefficient constrained metric learning (NCML), are developed. Both PCML and NCML can guarantee the global optimality of their solutions. Experiments are conducted on general classification, face verification, and person re-identification to evaluate our methods. Compared with the state-of-the-art approaches, our methods can achieve comparable classification accuracy and are efficient in training.

Single-image super-resolution (SR), which refers to reconstructing a higher resolution image from the observed lowresolution (LR) image, has received substantial attention due to its tremendous application potentials. Despite the breakthroughs of recently proposed SR methods using convolutional neural networks, their generated results usually lack of preserving structural (high-frequency) details. In this paper, regarding global boundary context and residual context as complimentary information for enhancing structural details in image restoration, we develop a contextualized multitask learning framework to address the SR problem. Specifically, our method first extracts convolutional features from the input LR image and applies one deconvolutional module to interpolate the LR feature maps in a content-adaptive way. Then, the resulting feature maps are fed into two branched subnetworks. On several standard benchmarks (e.g., Set5, Set14, and BSD200), our extensive evaluations demonstrate the effectiveness of our SR method on achieving both higher restoration quality and computational efficiency compared with several state-of-the-art SR approaches.

Saliency detection has recently received increasing research interest on using high-dimensional datasets beyond two-dimensional images. Despite the many available capturing devices and algorithms, there still exists a wide spectrum of challenges that need to be addressed to achieve accurate saliency detection. Inspired by the success of the light-field technique, in this article, we propose a new computational scheme to detect salient regions by integrating multiple visual cues from light-field images. First, saliency prior maps are generated from several light-field features based on superpixel-level intra-cue distinctiveness, such as color, depth, and flow inherited from different focal planes and multiple viewpoints. Then, we introduce the location prior to enhance the saliency maps. These maps will finally be merged into a single map using a random-search-based weighting strategy. Besides, we refine the object details by employing a two-stage saliency refinement to obtain the final saliency map. In addition, we present a more challenging benchmark dataset for light-field saliency analysis, named HFUT-Lytro , which consists of 255 light fields with a range from 53 to 64 images generated from each light-field image, therein spanning multiple occurrences of saliency detection challenges such as occlusions, cluttered background, and appearance changes. Experimental results show that our approach can achieve 0.6--6.7% relative improvements over state-of-the-art methods in terms of the F-measure and Precision metrics, which demonstrates the effectiveness of the proposed approach.

The past decade has witnessed the rapid development of feature representation learning and distance metric learning, whereas the two steps are often discussed separately. To explore their interaction, this work proposes an end-to-end learning framework called DARI, i.e. Distance metric And Representation Integration, and validates the effectiveness of DARI in the challenging task of person verification. Given the training images annotated with the labels, we first produce a large number of triplet units, and each one contains three images, i.e. one person and the matched/mismatch references. For each triplet unit, the distance disparity between the matched pair and the mismatched pair tends to be maximized. We solve this objective by building a deep architecture of convolutional neural networks. In particular, the Mahalanobis distance matrix is naturally factorized as one top fully-connected layer that is seamlessly integrated with other bottom layers representing the image feature. The image feature and the distance metric can be thus simultaneously optimized via the one-shot backward propagation. On several public datasets, DARI shows very promising performance on re-identifying individuals cross cameras against various challenges, and outperforms other state-of-the-art approaches.

Semantic object parsing is a fundamental task for understanding objects in detail in computer vision community, where incorporating multi-level contextual information is critical for achieving such fine-grained pixel-level recognition. Prior methods often leverage the contextual information through post-processing predicted confidence maps. In this work, we propose a novel deep Local-Global Long Short-Term Memory (LG-LSTM) architecture to seamlessly incorporate short-distance and long-distance spatial dependencies into the feature learning over all pixel positions. In each LG-LSTM layer, local guidance from neighboring positions and global guidance from the whole image are imposed on each position to better exploit complex local and global contextual information. Individual LSTMs for distinct spatial dimensions are also utilized to intrinsically capture various spatial layouts of semantic parts in the images, yielding distinct hidden and memory cells of each position for each dimension. In our parsing approach, several LG-LSTM layers are stacked and appended to the intermediate convolutional layers to directly enhance visual features, allowing network parameters to be learned in an end-to-end way. The long chains of sequential computation by stacked LG-LSTM layers also enable each pixel to sense a much larger region for inference benefiting from the memorization of previous dependencies in all positions along all dimensions. Comprehensive evaluations on three public datasets well demonstrate the significant superiority of our LG-LSTM over other state-of-the-art methods.

In this work, we propose a novel Reversible Recursive Instance-level Object Segmentation (R2-IOS) framework to address the challenging instance-level object segmentation task. R2-IOS consists of a reversible proposal refinement sub-network that predicts bounding box offsets for refining the object proposal locations, and an instance-level segmentation sub-network that generates the foreground mask of the dominant object instance in each proposal. By being recursive, R2-IOS iteratively optimizes the two subnetworks during joint training, in which the refined object proposals and improved segmentation predictions are alternately fed into each other to progressively increase the network capabilities. By being reversible, the proposal refinement sub-network adaptively determines an optimal number of refinement iterations required for each proposal during both training and testing. Furthermore, to handle multiple overlapped instances within a proposal, an instance-aware denoising autoencoder is introduced into the segmentation sub-network to distinguish the dominant object from other distracting instances. Extensive experiments on the challenging PASCAL VOC 2012 benchmark well demonstrate the superiority of R2-IOS over other state-of-the-art methods. In particular, the APr over 20 classes at 0:5 IoU achieves 66:7%, which significantly outperforms the results of 58:7% by PFN [17] and 46:3% by [22].

This paper addresses a fundamental problem of scene understanding: How to parse the scene image into a structured configuration (i.e., a semantic object hierarchy with object interaction relations) that finely accords with human perception. We propose a deep architecture consisting of two networks: i) a convolutional neural network (CNN) extracting the image representation for pixelwise object labeling and ii) a recursive neural network (RNN) discovering the hierarchical object structure and the inter-object relations. Rather than relying on elaborative user annotations (e.g., manually labeling semantic maps and relations), we train our deep model in a weakly-supervised manner by leveraging the descriptive sentences of the training images. Specifically, we decompose each sentence into a semantic tree consisting of nouns and verb phrases, and facilitate these trees discovering the configurations of the training images. Once these scene configurations are determined, then the parameters of both the CNN and RNN are updated accordingly by back propagation. The entire model training is accomplished through an Expectation-Maximization method. Extensive experiments suggest that our model is capable of producing meaningful and structured scene configurations and achieving more favorable scene labeling performance on PASCAL VOC 2012 over other state-of-the-art weakly-supervised methods.

Feature representation and object category classification are two key components of most object detection methods. While significant improvements have been achieved for deep feature representation learning, traditional SVM/softmax classifiers remain the dominant methods for the final object category classification. However, SVM/softmax classifiers lack the capacity of explicitly exploiting the complex structure of deep features, as they are purely discriminative methods. The recently proposed discriminative dictionary pair learning (DPL) model involves a fidelity term to minimize the reconstruction loss and a discrimination term to enhance the discriminative capability of the learned dictionary pair, and thus is appropriate for balancing the representation and discrimination to boost object detection performance. In this paper, we propose a novel object detection system by unifying DPL with the convolutional feature learning. Specifically, we incorporate DPL as a Dictionary Pair Classifier Layer (DPCL) into the deep architecture, and develop an end-to-end learning algorithm for optimizing the dictionary pairs and the neural networks simultaneously. Moreover, we design a multi-task loss for guiding our model to accomplish the three correlated tasks: objectness estimation, categoryness computation, and bounding box regression. From the extensive experiments on PASCAL VOC 2007/2012 benchmarks, our approach demonstrates the effectiveness to substantially improve the performances over the popular existing object detection frameworks (e.g., R-CNN [13] and FRCN [12]), and achieves new state-of-the-arts.

Person re-identification has been usually solved as either the matching of single-image representation (SIR) or the classification of cross-image representation (CIR). In this work, we exploit the connection between these two categories of methods, and propose a joint learning frame-work to unify SIR and CIR using convolutional neural network (CNN). Specifically, our deep architecture contains one shared sub-network together with two sub-networks that extract the SIRs of given images and the CIRs of given image pairs, respectively. The SIR sub-network is required to be computed once for each image (in both the probe and gallery sets), and the depth of the CIR sub-network is required to be minimal to reduce computational burden. Therefore, the two types of representation can be jointly optimized for pursuing better matching accuracy with moderate computational cost. Furthermore, the representations learned with pairwise comparison and triplet comparison objectives can be combined to improve matching performance. Experiments on the CUHK03, CUHK01 and VIPeR datasets show that the proposed method can achieve favorable accuracy while compared with state-of-the-arts.

Maximally stable extremal regions (MSER), which is a popular method to generate character proposals/candidates, has shown superior performance in scene text detection. However, the pixel-level operation limits its capability for handling some challenging cases (e.g., multiple connected characters, separated parts of one character and non-uniform illumination). To better tackle these cases, we design a character proposal network (CPN) by taking advantage of the high capacity and fast computing of fully convolutional network (FCN). Specifically, the network simultaneously predicts character-ness scores and refines the corresponding locations. The character-ness scores can be used for proposal ranking to reject non-character proposals and the refining process aims to obtain the more accurate locations. Furthermore, considering the situation that different characters have different aspect ratios, we propose a multi-template strategy, designing a refiner for each aspect ratio. The extensive experiments indicate our method achieves recall rates of 93.88%, 93.60% and 96.46% on ICDAR 2013, SVT and Chinese 2k datasets respectively using less than 1000 proposals, demonstrating promising performance of our character proposal network.

Recently, machine learning based single image super resolution (SR) approaches focus on jointly learning representations for high-resolution (HR) and low-resolution (LR) image patch pairs to improve the quality of the super-resolved images. However, due to treat all image pixels equally without considering the salient structures, these approaches usually fail to produce visual pleasant images with sharp edges and fine details. To address this issue, in this work we present a new novel SR approach, which replaces the main building blocks of the classical interpolation pipeline by a flexible, content-adaptive deep neural networks. In particular, two well-designed structure-aware components, respectively capturing local- and holistic-image contents, are naturally incorporated into the fully-convolutional representation learning to enhance the image sharpness and naturalness. Extensively evaluations on several standard benchmarks (e.g., Set5, Set14 and BSD200) demonstrate that our approach can achieve superior results, especially on the image with salient structures, over many existing state-of-the-art SR methods under both quantitative and qualitative measures.

This paper proposes a lightweight deep model to recognize age and gender from a face image. Though simple, our network architecture is able to complete the two tasks effectively and efficiently. Moreover, different from existing methods, we simultaneously perform the age and gender recognition tasks via a joint regression model. Specifically, our model employs a multi-task learning scheme to learn shared features for these two correlated tasks in an end-to-end manner. Extensive experimental results on the recent Adience benchmark demonstrate that our model achieves competitive recognition accuracy with the state-of-the-art methods but with much faster speed, i.e., about 10 times faster in the testing phase. Our model can be easily adopted and extended to other facial applications.

Human pose estimation (i.e., locating the body parts / joints of a person) is a fundamental problem in human-computer interaction and multimedia applications. Significant progress has been made based on the development of depth sensors, i.e., accessible human pose prediction from still depth images~\cite{rf12pami}. However, most of the existing approaches to this problem involve several components/models that are independently designed and optimized, leading to suboptimal performances. In this paper, we propose a novel inference-embedded multi-task learning framework for predicting human pose from still depth images, which is implemented with a deep architecture of neural networks. Specifically, we handle two cascaded tasks: i) generating the heat (confidence) maps of body parts via a fully convolutional network (FCN); ii) seeking the optimal configuration of body parts based on the detected body part proposals via an inference built-in MatchNet~\cite{mn15cvpr}, which measures the appearance and geometric kinematic compatibility of body parts and embodies the dynamic programming inference as an extra network layer. These two tasks are jointly optimized. Our extensive experiments show that the proposed deep model significantly improves the accuracy of human pose estimation over other several state-of-the-art methods or SDKs. We also release a large-scale dataset for comparison, which includes 100K depth images under challenging scenarios.

This paper presents a conceptually simple but effective approach to track multiobject in videos without requiring elaborate supervision (i.e., training object detectors or templates offline). Our framework performs a bi-layer inference of spatio-temporal grouping to exploit rich appearance and motion information in the observed sequence. First, we generate a robust middle-level video representation based on clustered point tracks, namely video bundles. Each bundle encapsulates a chunk of point tracks satisfying both spatial proximity and temporal coherency. Taking the video bundles as vertices, we build a spatio-temporal graph that incorporates both competitive and compatible relations among vertices. The multiobject tracking can be then phrased as a graph partition problem under the Bayesian framework, and we solve it by developing a reconfigurable belief propagation (BP) algorithm. This algorithm improves the traditional BP method by allowing a converged solution to be reconfigured during optimization, so that the inference can be reactivated once it gets stuck in local minima and thus conduct more reliable results. In the experiments, we demonstrate the superior performances of our approach on the challenging benchmarks compared with other state-of-the-art methods.

Segmenting organisms or tumors from medical data (e.g., computed tomography volumetric images, ultrasound, or magnetic resonance imaging images/image sequences) is one of the fundamental tasks in medical image analysis and diagnosis, and has received long-term attentions. This paper studies a novel computational framework of interactive segmentation for extracting liver tumors from image sequences, and it is suitable for different types of medical data. The main contributions are twofold. First, we propose a collaborative model to jointly formulate the tumor segmentation from two aspects: 1) region partition and 2) boundary presence. The two terms are complementary but simultaneously competing: the former extracts the tumor based on its appearance/texture information, while the latter searches for the palpable tumor boundary. Moreover, in order to adapt the data variations, we allow the model to be discriminatively trained based on both the seed pixels traced by the Lucas-Kanade algorithm and the scribbles placed by the user. Second, we present an effective inference algorithm that iterates to: 1) solve tumor segmentation using the augmented Lagrangian method and 2) propagate the segmentation across the image sequence by searching for distinctive matches between images. We keep the collaborative model updated during the inference in order to well capture the tumor variations over time. We have verified our system for segmenting liver tumors from a number of clinical data, and have achieved very promising results. The software developed with this paper can be found at http://vision.sysu.edu.cn/projects/med-interactive-seg/.

Integrating multiple different yet complementary feature representations has been proved to be an effective way for boosting tracking performance. This paper investigates how to perform robust object tracking in challenging scenarios by adaptively incorporating information from grayscale and thermal videos, and proposes a novel collaborative algorithm for online tracking. In particular, an adaptive fusion scheme is proposed based on collaborative sparse representation in Bayesian filtering framework. We jointly optimize sparse codes and the reliable weights of different modalities in an online way. In addition, this paper contributes a comprehensive video benchmark, which includes 50 grayscale-thermal sequences and their ground truth annotations for tracking purpose. The videos are with high diversity and the annotations were finished by one single person to guarantee consistency. Extensive experiments against other state-of-the-art trackers with both grayscale and grayscale-thermal inputs demonstrate the effectiveness of the proposed tracking approach. Through analyzing quantitative results, we also provide basic insights and potential future research directions in grayscale-thermal tracking.

This paper investigates how to perform robust and efficient video segmentation while suppressing the effects of data noises and/or corruptions, and an effective approach is introduced to this end. First, a general algorithm, called sub-optimal low-rank decomposition (SOLD), is proposed to pursue the low-rank representation for video segmentation. Given the data matrix formed by supervoxel features of an observed video sequence, SOLD seeks a sub-optimal solution by making the matrix rank explicitly determined. In particular, the representation coefficient matrix with the fixed rank can be decomposed into two sub-matrices of low rank, and then we iteratively optimize them with closed-form solutions. Moreover, we incorporate a discriminative replication prior into SOLD based on the observation that small-size video patterns tend to recur frequently within the same object. Second, based on SOLD, we present an efficient inference algorithm to perform streaming video segmentation in both unsupervised and interactive scenarios. More specifically, the constrained normalized-cut algorithm is adopted by incorporating the low-rank representation with other low level cues and temporal consistent constraints for spatio-temporal segmentation. Extensive experiments on two public challenging data sets VSB100 and SegTrack suggest that our approach outperforms other video segmentation approaches in both accuracy and efficiency.

This work investigates how to automatically classify Focal Liver Lesions (FLLs) into three specific benign or malignant types in Contrast-Enhanced Ultrasound (CEUS) videos, and aims at providing a computational framework to assist clinicians in FLL diagnosis. The main challenge for this task is that FLLs in CEUS videos often show diverse enhancement patterns at different temporal phases. To handle these diverse patterns, we propose a novel structured model, which detects a number of discriminative Regions of Interest (ROIs) for the FLL and recognize the FLL based on these ROIs. Our model incorporates an ensemble of local classifiers in the attempt to identify different enhancement patterns of ROIs, and in particular, we make the model reconfigurable by introducing switch variables to adaptively select appropriate classifiers during inference. We formulate the model learning as a non-convex optimization problem, and present a principled optimization method to solve it in a dynamic manner: the latent structures (e.g. the selections of local classifiers, and the sizes and locations of ROIs) are iteratively determined along with the parameter learning. Given the updated model parameters in each step, the data-driven inference is also proposed to efficiently determine the latent structures by using the sequential pruning and dynamic programming method. In the experiments, we demonstrate superior performances over the state-of-the-art approaches. We also release hundreds of CEUS FLLs videos used to quantitatively evaluate this work, which to the best of our knowledge forms the largest dataset in the literature. Please find more information at "http://vision.sysu.edu.cn/projects/fllrecog/".

This paper aims at developing an integrated system for clothing co-parsing (CCP), in order to jointly parse a set of clothing images (unsegmented but annotated with tags) into semantic configurations. A novel data-driven system consisting of two phases of inference is proposed. The first phase, referred as “image cosegmentation,” iterates to extract consistent regions on images and jointly refines the regions over all images by employing the exemplar-SVM technique [1]. In the second phase (i.e., “region colabeling”), we construct a multiimage graphical model by taking the segmented regions as vertices, and incorporating several contexts of clothing configuration (e.g., item locations and mutual interactions). The joint label assignment can be solved using the efficient Graph Cuts algorithm. In addition to evaluate our framework on the Fashionista dataset [2], we construct a dataset called the SYSU-Clothes dataset consisting of 2098 high-resolution street fashion photos to demonstrate the performance of our system. We achieve 90.29%/88.23% segmentation accuracy and 65.52%/63.89% recognition rate on the Fashionista and the SYSU-Clothes datasets, respectively, which are superior compared with the previous methods. Furthermore, we apply our method on a challenging task, i.e., cross-domain clothing retrieval: given user photo depicting a clothing image, retrieving the same clothing items from online shopping stores based on the fine-grained parsing results.

Salient object detection increasingly receives attention as an important component or step in several pattern recognition and image processing tasks. Although a variety of powerful saliency models have been intensively proposed, they usually involve heavy feature (or model) engineering based on priors (or assumptions) about the properties of objects and backgrounds. Inspired by the effectiveness of recently developed feature learning, we provide a novel deep image saliency computing (DISC) framework for fine-grained image saliency computing. In particular, we model the image saliency from both the coarse-and fine-level observations, and utilize the deep convolutional neural network (CNN) to learn the saliency representation in a progressive manner. In particular, our saliency model is built upon two stacked CNNs. The first CNN generates a coarse-level saliency map by taking the overall image as the input, roughly identifying saliency regions in the global context. Furthermore, we integrate superpixel-based local context information in the first CNN to refine the coarse-level saliency map. Guided by the coarse saliency map, the second CNN focuses on the local context to produce fine-grained and accurate saliency map while preserving object details. For a testing image, the two CNNs collaboratively conduct the saliency computing in one shot. Our DISC framework is capable of uniformly highlighting the objects of interest from complex background while preserving well object details. Extensive experiments on several standard benchmarks suggest that DISC outperforms other state-of-the-art methods and it also generalizes well across data sets without additional training. The executable version of DISC is available online: http://vision.sysu.edu.cn/projects/DISC.

This paper presents a novel compositional contour-based shape model by incorporating multiple distance metrics to account for varying shape distortions or deformations. Our approach contains two key steps: 1) contour feature generation and 2) generative model pursuit. For each category, we first densely sample an ensemble of local prototype contour segments from a few positive shape examples and describe each segment using three different types of distance metrics. These metrics are diverse and complementary with each other to capture various shape deformations. We regard the parameterized contour segment plus an additive residual ϵ as a basic subspace, namely, ϵ -ball, in the sense that it represents local shape variance under the certain distance metric. Using these ϵ -balls as features, we then propose a generative learning algorithm to pursue the compositional shape model, which greedily selects the most representative features under the information projection principle. In experiments, we evaluate our model on several public challenging data sets, and demonstrate that the integration of multiple shape distance metrics is capable of dealing various shape deformations, articulations, and background clutter, hence boosting system performance.

This paper investigates how to perform robust and efficient unsupervised video segmentation while suppressing the effects of data noises and/or corruptions. We propose a general algorithm, called Sub-Optimal Low-rank Decomposition (SOLD), which pursues the low-rank representation for video segmentation. Given the supervoxels affinity matrix of an observed video sequence, SOLD seeks a sub-optimal solution by making the matrix rank explicitly determined. In particular, the affinity matrix with the rank fixed can be decomposed into two sub-matrices of low rank, and then we iteratively optimize them with closed-form solutions. Moreover, we incorporate a discriminative replication prior into our framework based on the obervation that small-size video patterns tend to recur frequently within the same object. The video can be segmented into several spatio-temporal regions by applying the Normalized-Cut (NCut) algorithm with the solved low-rank representation. To process the streaming videos, we apply our algorithm sequentially over a batch of frames over time, in which we also develop several temporal consistent constraints improving the robustness. Extensive experiments on the public benchmarks demonstrate superior performance of our framework over other state-of-the-art approaches.

Both parametric and non-parametric approaches have demonstrated encouraging performances in the human parsing task, namely segmenting a human image into several semantic regions (e.g., hat, bag, left arm, face). In this work, we aim to develop a new solution with the advantages of both methodologies, namely supervision from annotated data and the flexibility to use newly annotated (possibly uncommon) images, and present a quasi-parametric human parsing model. Under the classic K Nearest Neighbor (KNN)-based nonparametric framework, the parametric Matching Convolutional Neural Network (M-CNN) is proposed to predict the matching confidence and displacements of the best matched region in the testing image for a particular semantic region in one KNN image. Given a testing image, we first retrieve its KNN images from the annotated/manually-parsed human image corpus. Then each semantic region in each KNN image is matched with confidence to the testing image using M-CNN, and the matched regions from all KNN images are further fused, followed by a superpixel smoothing procedure to obtain the ultimate human parsing result. The M-CNN differs from the classic CNN [12] in that the tailored cross image matching filters are introduced to characterize the matching between the testing image and the semantic region of a KNN image. The cross image matching filters are defined at different convolutional layers, each aiming to capture a particular range of displacements. Comprehensive evaluations over a large dataset with 7,700 annotated human images well demonstrate the significant performance gain from the quasi-parametric model over the state-of-the-arts [29, 30], for the human parsing task.

The maximum a posterior (MAP)-based blind deconvolution framework generally involves two stages: blur kernel estimation and non-blind restoration. For blur kernel estimation, sharp edge prediction and carefully designed image priors are vital to the success of MAP. In this paper, we propose a blind deconvolution framework together with iteration specific priors for better blur kernel estimation. The family of hyper-Laplacian (Pr(d) ∝ e-∥d∥pp/λ) is adopted for modeling iteration-wise priors of image gra- dients, where each iteration has its own model parameters {λ(t), p(t)}. To avoid heavy parameter tuning, all iteration-wise model parameters can be learned using our principled discriminative learning model from a training set, and can be directly applied to other dataset and real blurry images. Interestingly, with the generalized shrinkage / thresholding operator, negative p value (p <;0) is allowable and we find that it contributes more in estimating the coarse shape of blur kernel. Experimental results on synthetic and real world images demonstrate that our method achieves better deblurring results than the existing gradient prior-based methods. Compared with the state-of-the-art patch prior-based method, our method is competitive in restoration results but is much more efficient.

Intuitive observations show that a baby may inherently possess the capability of recognizing a new visual concept (e.g., chair, dog) by learning from only very few positive instances taught by parent(s) or others, and this recognition capability can be gradually further improved by exploring and/or interacting with the real instances in the physical world. Inspired by these observations, we propose a computational model for weakly-supervised object detection, based on prior knowledge modelling, exemplar learning and learning with video contexts. The prior knowledge is modeled with a pre-trained Convolutional Neural Network (CNN). When very few instances of a new concept are given, an initial concept detector is built by exemplar learning over the deep features the pre-trained CNN. The well-designed tracking solution is then used to discover more diverse instances from the massive online weakly labeled videos. Once a positive instance is detected/identified with high score in each video, more instances possibly from different view-angles and/or different distances are tracked and accumulated. Then the concept detector can be fine-tuned based on these new instances. This process can be repeated again and again till we obtain a very mature concept detector. Extensive experiments on Pascal VOC-07/10/12 object detection datasets [9] well demonstrate the effectiveness of our framework. It can beat the state-of-the-art full-training based performances by learning from very few samples for each object category, along with about 20,000 weakly labeled videos.

In this work, we address the human parsing task with a novel Contextualized Convolutional Neural Network (Co-CNN) architecture, which well integrates the cross-layer context, global image-level context, within-super-pixel context and cross-super-pixel neighborhood context into a unified network. Given an input human image, Co-CNN produces the pixel-wise categorization in an end-to-end way. First, the cross-layer context is captured by our basic local-to-global-to-local structure, which hierarchically combines the global semantic structure and the local fine details within the cross-layers. Second, the global image-level label prediction is used as an auxiliary objective in the intermediate layer of the Co-CNN, and its outputs are further used for guiding the feature learning in subsequent convolutional layers to leverage the global image-level context. Finally, to further utilize the local super-pixel contexts, the within-super-pixel smoothing and cross-super-pixel neighbourhood voting are formulated as natural sub-components of the Co-CNN to achieve the local label consistency in both training and testing process. Comprehensive evaluations on two public datasets well demonstrate the significant superiority of our Co-CNN architecture over other state-of-the-arts for human parsing. In particular, the F-1 score on the large dataset [15] reaches 76.95% by Co-CNN, significantly higher than 62.81% and 64.38% by the state-of-the-art algorithms, M-CNN [21] and ATR [15], respectively.

Sketch-based face recognition is an interesting task in vision and multimedia research, yet it is quite challenging due to the great difference between face photos and sketches. In this paper, we propose a novel approach for photo-sketch generation, aiming to automatically transform face photos into detail-preserving personal sketches. Unlike the traditional models synthesizing sketches based on a dictionary of exemplars, we develop a fully convolutional network to learn the end-to-end photo-sketch mapping. Our approach takes whole face photos as inputs and directly generates the corresponding sketch images with efficient inference and learning, in which the architecture is stacked by only convolutional kernels of very small sizes. To well capture the person identity during the photo-sketch transformation, we define our optimization objective in the form of joint generative discriminative minimization. In particular, a discriminative regularization term is incorporated into the photo-sketch generation, enhancing the discriminability of the generated person sketches against other individuals. Extensive experiments on several standard benchmarks suggest that our approach outperforms other state-of-the-arts in both photo sketch generation and face sketch verification.

This article summarizes the corresponding half-day tutorial at ACM Multimedia 2015. This tutorial reviews recent progresses in human-centric images and videos analysis: 1) fashion analysis: parsing, attribute prediction and retrieval; 2) action analysis: discriminative feature selection, pooling and fusion; 3) person verification: cross-domain person verification via learning a generalized similarity measure, and bit-scalable deep hashing with regularized similarity learning.

In this work, the human parsing task, namely decomposing a human image into semantic fashion/body regions, is formulated as an active template regression (ATR) problem, where the normalized mask of each fashion/body item is expressed as the linear combination of the learned mask templates, and then morphed to a more precise mask with the active shape parameters, including position, scale and visibility of each semantic region. The mask template coefficients and the active shape parameters together can generate the human parsing results, and are thus called the structure outputs for human parsing. The deep Convolutional Neural Network (CNN) is utilized to build the end-to-end relation between the input human image and the structure outputs for human parsing. More specifically, the structure outputs are predicted by two separate networks. The first CNN network is with max-pooling, and designed to predict the template coefficients for each label mask, while the second CNN network is without max-pooling to preserve sensitivity to label mask position and accurately predict the active shape parameters. For a new image, the structure outputs of the two networks are fused to generate the probability of each label for each pixel, and super-pixel smoothing is finally used to refine the human parsing result. Comprehensive evaluations on a large dataset well demonstrate the significant superiority of the ATR framework over other state-of-the-arts for human parsing. In particular, the F1-score reaches 64.38 percent by our ATR framework, significantly higher than 44.76 percent based on the state-of-the-art algorithm [28].

In this paper, we investigate a novel reconfigurable part-based model, namely And-Or graph model, to recognize object shapes in images. Our proposed model consists of four layers: leaf-nodes at the bottom are local classifiers for detecting contour fragments; or-nodes above the leaf-nodes function as the switches to activate their child leaf-nodes, making the model reconfigurable during inference; and-nodes in a higher layer capture holistic shape deformations; one root-node on the top, which is also an or-node, activates one of its child and-nodes to deal with large global variations (e.g. different poses and views). We propose a novel structural optimization algorithm to discriminatively train the And-Or model from weakly annotated data. This algorithm iteratively determines the model structures (e.g. the nodes and their layouts) along with the parameter learning. On several challenging datasets, our model demonstrates the effectiveness to perform robust shape-based object detection against background clutter and outperforms the other state-of-the-art approaches. We also release a new shape database with annotations, which includes more than 1500 challenging shape instances, for recognition and detection.

This paper investigates a general framework to discover categories of unlabeled scene images according to their appearances (i.e., textures and structures). We jointly solve the two coupled tasks in an unsupervised manner: 1) classifying images without predetermining the number of categories and 2) pursuing generative model for each category. In our method, each image is represented by two types of image descriptors that are effective to capture image appearances from different aspects. By treating each image as a graph vertex, we build up a graph and pose the image categorization as a graph partition process. Specifically, a partitioned subgraph can be regarded as a category of scenes and we define the probabilistic model of graph partition by accumulating the generative models of all separated categories. For efficient inference with the graph, we employ a stochastic cluster sampling algorithm, which is designed based on the Metropolis-Hasting mechanism. During the iterations of inference, the model of each category is analytically updated by a generative learning algorithm. In the experiments, our approach is validated on several challenging databases, and it outperforms other popular state-of-the-art methods. The implementation details and empirical analysis are presented as well.

Driven by recent vision and graphics applications such as image segmentation and object recognition, computing pixel-accurate saliency values to uniformly highlight foreground objects becomes increasingly important. In this paper, we propose a unified framework called pixelwise image saliency aggregating (PISA) various bottom-up cues and priors. It generates spatially coherent yet detail-preserving, pixel-accurate, and fine-grained saliency, and overcomes the limitations of previous methods, which use homogeneous superpixel based and color only treatment. PISA aggregates multiple saliency cues in a global context, such as complementary color and structure contrast measures, with their spatial priors in the image domain. The saliency confidence is further jointly modeled with a neighborhood consistence constraint into an energy minimization formulation, in which each pixel will be evaluated with multiple hypothetical saliency levels. Instead of using global discrete optimization methods, we employ the cost-volume filtering technique to solve our formulation, assigning the saliency levels smoothly while preserving the edge-aware structure details. In addition, a faster version of PISA is developed using a gradient-driven image subsampling strategy to greatly improve the runtime efficiency while keeping comparable detection accuracy. Extensive experiments on a number of public data sets suggest that PISA convincingly outperforms other state-of-the-art approaches. In addition, with this work, we also create a new data set containing 800 commodity images for evaluating saliency detection.

Extracting informative image features and learning effective approximate hashing functions are two crucial steps in image retrieval. Conventional methods often study these two steps separately, e.g., learning hash functions from a predefined hand-crafted feature space. Meanwhile, the bit lengths of output hashing codes are preset in the most previous methods, neglecting the significance level of different bits and restricting their practical flexibility. To address these issues, we propose a supervised learning framework to generate compact and bit-scalable hashing codes directly from raw images. We pose hashing learning as a problem of regularized similarity learning. In particular, we organize the training images into a batch of triplet samples, each sample containing two images with the same label and one with a different label. With these triplet samples, we maximize the margin between the matched pairs and the mismatched pairs in the Hamming space. In addition, a regularization term is introduced to enforce the adjacency consistency, i.e., images of similar appearances should have similar codes. The deep convolutional neural network is utilized to train the model in an end-to-end fashion, where discriminative image features and hash functions are simultaneously optimized. Furthermore, each bit of our hashing codes is unequally weighted, so that we can manipulate the code lengths by truncating the insignificant bits. Our framework outperforms state-of-the-arts on public benchmarks of similar image search and also achieves promising results in the application of person re-identification in surveillance. It is also shown that the generated bit-scalable hashing codes well preserve the discriminative powers with shorter code lengths.

In this paper, we propose a novel semi- supervised learning strategy to address human parsing. Existing human parsing datasets are relatively small due to the required tedious human labeling. We present a general, affordable and scalable solution, which harnesses the rich contexts in those easily available web videos to boost any existing human parser. First, we crawl a large number of unlabeled videos from the web. Then for each video, the cross-frame contexts are utilized for human pose co- estimation , and then video co-parsing to obtain satisfactory human parsing results for all frames. More specifically, SIFT flow and super-pixel matching are used to build correspondences across different frames, and these correspondences then contextualize the pose estimation and human parsing in individual frames. Finally these parsed video frames are used as the reference corpus for the non-parametric human parsing component of the whole solution. To further improve the accuracy of video co-parsing, we propose an active learning method to incorporate human guidance, where the labelers are required to assess the accuracies of the pose estimation results of certain selected video frames. Then we take reliable frames as the seed frames to guide the video pose co-estimation. Our human parsing framework can then easily incorporate the human feedback to train a better fashion parser. Extensive experiments on two benchmark fashion datasets as well as a newly collected challenging Fashion Icon dataset well demonstrate the encouraging performance gain from our general pipeline for human parsing.

This paper aims at developing an integrated system of clothing co-parsing, in order to jointly parse a set of clothing images (unsegmented but annotated with tags) into semantic configurations. We propose a data-driven framework consisting of two phases of inference. The first phase, referred as "image co-segmentation", iterates to extract consistent regions on images and jointly refines the regions over all images by employing the exemplar-SVM (ESVM) technique [23]. In the second phase (i.e. "region colabeling"), we construct a multi-image graphical model by taking the segmented regions as vertices, and incorporate several contexts of clothing configuration (e.g., item location and mutual interactions). The joint label assignment can be solved using the efficient Graph Cuts algorithm. In addition to evaluate our framework on the Fashionista dataset [30], we construct a dataset called CCP consisting of 2098 high-resolution street fashion photos to demonstrate the performance of our system. We achieve 90.29% / 88.23% segmentation accuracy and 65.52% / 63.89% recognition rate on the Fashionista and the CCP datasets, respectively, which are superior compared with state-of-the-art methods.

This paper aims at one newly raising task in vision and multimedia research: recognizing human actions from still images. Its main challenges lie in the large variations in human poses and appearances, as well as the lack of temporal motion information. Addressing these problems, we propose to develop an expressive deep model to naturally integrate human layout and surrounding contexts for higher level action understanding from still images. In particular, a Deep Belief Net is trained to fuse information from different noisy sources such as body part detection and object detection. To bridge the semantic gap, we used manually labeled data to greatly improve the effectiveness and efficiency of the pre-training and fine-tuning stages of the DBN training. The resulting framework is shown to be robust to sometimes unreliable inputs (e.g., imprecise detections of human parts and objects), and outperforms the state-of-the-art approaches.

This article studies a data-driven approach for semantically scene understanding, without pixelwise annotation and classifier pre-training. Our framework parses a target image with two steps: (i) retrieving its exemplars (i.e. references) from an image database, where all images are unsegmented but annotated with tags; (ii) recovering its pixel labels by propagating semantics from the references. We present a novel framework making the two steps mutually conditional and bootstrapped under the probabilistic Expectation-maximization (EM) formulation. In the first step, the references are selected by jointly matching their appearances with the target as well as the semantics. We process the second step via a combinatorial graphical representation, in which the vertices are superpixels extracted from the target and its selected references. Then we derive the potentials of assigning labels to one vertex of the target, which depends upon the graph edges that connect the vertex to its spatial neighbors of the target and to its similar vertices of the references. Two steps can be both solved analytically, and the inference is conducted in a self-driven fashion. In the experiments, we validate our approach on two public databases, and demonstrate superior performances over the state-of-the-art methods.

Constructing effective representations is a critical but challenging problem in multimedia understanding. The traditional handcraft features often rely on domain knowledge, limiting the performances of exiting methods. This paper discusses a novel computational architecture for general image feature mining, which assembles the primitive filters (i.e. Gabor wavelets) into compositional features in a layer-wise manner. In each layer, we produce a number of base classifiers (i.e. regression stumps) associated with the generated features, and discover informative compositions by using the boosting algorithm. The output compositional features of each layer are treated as the base components to build up the next layer. Our framework is able to generate expressive image representations while inducing very discriminate functions for image classification. The experiments are conducted on several public datasets, and we demonstrate superior performances over state-of-the-art approaches.

In this paper, we explore how to utilize the video context to facilitate fashion parsing. Instead of annotating a large amount of fashion images, we present a general, affordable and scalable solution, which harnesses the rich contexts in easily available fashion videos to boost any existing fashion parser. First, we crawl a large unlabelled fashion video corpus with fashion frames. Then for each fashion video, the cross-frame contexts are utilized for human pose co-estimation, and then video co-parsing to obtain satisfactory fashion parsing results for all frames. More specifically, Sift Flow and super-pixel matching are used to build correspondences across frames, and these correspondences then con- textualize the pose estimations and fashion parsing in individual frames. Finally, these parsed video frames are used as the reference corpus for the non-parametric fashion parsing component of the whole solution. Extensive experiments on two benchmark fashion datasets as well as a newly collected challenging Fashion Icon (FI) dataset demonstrate the encouraging performance gain from our general pipeline for fashion parsing.

Human activity understanding with 3D/depth sensors has received increasing attention in multimedia processing and interactions. This work targets on developing a novel deep model for automatic activity recognition from RGB-D videos. We represent each human activity as an ensemble of cubic-like video segments, and learn to discover the temporal structures for a category of activities, i.e. how the activities to be decomposed in terms of classification. Our model can be regarded as a structured deep architecture, as it extends the convolutional neural networks (CNNs) by incorporating structure alternatives. Specifically, we build the network consisting of 3D convolutions and max-pooling operators over the video segments, and introduce the latent variables in each convolutional layer manipulating the activation of neurons. Our model thus advances existing approaches in two aspects: (i) it acts directly on the raw inputs (grayscale-depth data) to conduct recognition instead of relying on hand-crafted features, and (ii) the model structure can be dynamically adjusted accounting for the temporal variations of human activities, i.e. the network configuration is allowed to be partially activated during inference. For model training, we propose an EM-type optimization method that iteratively (i) discovers the latent structure by determining the decomposed actions for each training example, and (ii) learns the network parameters by using the back-propagation algorithm. Our approach is validated in challenging scenarios, and outperforms state-of-the-art methods. A large human activity database of RGB-D videos is presented in addition.

This paper presents a novel framework for a multimedia search task: searching a person in a scene using human body appearance. Existing works mostly focus on two independent problems related to this task, i.e., people detection and person re-identification. However, a sequential combination of these two components does not solve the person search problem seamlessly for two reasons: 1) the errors in people detection are carried into person re-identification unavoidably; 2) the setting of person re-identification is different from that of person search which is essentially a verification problem. To bridge this gap, we propose a unified framework which jointly models the commonness of people (for detection) and the uniqueness of a person (for identification). We demonstrate superior performance of our approach on public benchmarks compared with the sequential combination of the state-of-the-art detection and identification algorithms.

Feature point matching that incorporates pairwise constraints can be cast as an integer quadratic programming (IQP) problem. Since it is NP-hard, approximate methods are required. The optimal solution for IQP matching problem is discrete, binary, and thus sparse in nature. This motivates us to use sparse model for feature point matching problem. The main advantage of the proposed sparse feature point matching (SPM) method is that it generates sparse solution and thus naturally imposes the discrete mapping constraints approximately in the optimization process. Therefore, it can optimize the IQP matching problem in an approximate discrete domain. In addition, an efficient algorithm can be derived to solve SPM problem. Promising experimental results on both synthetic points sets matching and real-world image feature sets matching tasks show the effectiveness of the proposed feature point matching method.

Although it has been widely discussed in video surveillance, background subtraction is still an open problem in the context of complex scenarios, e.g., dynamic backgrounds, illumination variations, and indistinct foreground objects. To address these challenges, we propose an effective background subtraction method by learning and maintaining an array of dynamic texture models within the spatio-temporal representations. At any location of the scene, we extract a sequence of regular video bricks, i.e., video volumes spanning over both spatial and temporal domain. The background modeling is thus posed as pursuing subspaces within the video bricks while adapting the scene variations. For each sequence of video bricks, we pursue the subspace by employing the auto regressive moving average model that jointly characterizes the appearance consistency and temporal coherence of the observations. During online processing, we incrementally update the subspaces to cope with disturbances from foreground objects and scene changes. In the experiments, we validate the proposed method in several complex scenarios, and show superior performances over other state-of-the-art approaches of background subtraction. The empirical studies of parameter setting and component analysis are presented as well.

In this work, we present an efficient multi-scale low-rank representation for image segmentation. Our method begins with partitioning the input images into a set of super pixels, followed by seeking the optimal super pixel-pair affinity matrix, both of which are performed at multiple scales of the input images. Since low-level super pixel features are usually corrupted by image noises, we propose to infer the low-rank refined affinity matrix. The inference is guided by two observations on natural images. First, looking into a single image, local small-size image patterns tend to recur frequently within the same semantic region, but may not appear in semantically different regions. We call this internal image statistics as replication prior, and quantitatively justify it on real image databases. Second, the affinity matrices at different scales should be consistently solved, which leads to the cross-scale consistency constraint. We formulate these two purposes with one unified formulation and develop an efficient optimization procedure. Our experiments demonstrate the presented method can substantially improve segmentation accuracy.

Driven by recent vision and graphics applications such as image segmentation and object recognition, assigning pixel-accurate saliency values to uniformly highlight foreground objects becomes increasingly critical. More often, such fine-grained saliency detection is also desired to have a fast runtime. Motivated by these, we propose a generic and fast computational framework called PISA - Pixel wise Image Saliency Aggregating complementary saliency cues based on color and structure contrasts with spatial priors holistically. Overcoming the limitations of previous methods often using homogeneous super pixel-based and color contrast-only treatment, our PISA approach directly performs saliency modeling for each individual pixel and makes use of densely overlapping, feature-adaptive observations for saliency measure computation. We further impose a spatial prior term on each of the two contrast measures, which constrains pixels rendered salient to be compact and also centered in image domain. By fusing complementary contrast measures in such a pixel wise adaptive manner, the detection effectiveness is significantly boosted. Without requiring reliable region segmentation or post-relaxation, PISA exploits an efficient edge-aware image representation and filtering technique and produces spatially coherent yet detail-preserving saliency maps. Extensive experiments on three public datasets demonstrate PISA's superior detection accuracy and competitive runtime speed over the state-of-the-arts approaches.

This paper proposes a reconfigurable model to recognize and detect multiclass (or multiview) objects with large variation in appearance. Compared with well acknowledged hierarchical models, we study two advanced capabilities in hierarchy for object modeling: (i)``switch'' variables(i.e. or-nodes) for specifying alternative compositions, and (ii) making local classifiers (i.e. leaf-nodes) shared among different classes. These capabilities enable us to account well for structural variabilities while preserving the model compact. Our model, in the form of an And-Or Graph, comprises four layers: a batch of leaf-nodes with collaborative edges in bottom for localizing object parts, the or-nodes over bottom to activate their children leaf-nodes, the and-nodes to classify objects as a whole, one root-node on the top for switching multiclass classification, which is also an or-node. For model training, we present an EM-type algorithm, namely dynamical structural optimization (DSO), to iteratively determine the structural configuration, (e.g., leaf-node generation associated with their parent or-nodes and shared across other classes), along with optimizing multi-layer parameters. The proposed method is valid on challenging databases, e.g., PASCAL VOC 2007 and UIUC-People, and it achieves state-of-the-arts performance.

Recently, studies on sketch, such as sketch retrieval and sketch classification, have received more attention in the computer vision community. One of its most fundamental and essential problems is how to more effectively describe a sketch image. Many existing descriptors, such as shape context, have achieved great success. In this paper, we propose a new descriptor, namely Symmetric-aware Flip Invariant Sketch Histogram (SYM-FISH) to refine the shape context feature. Its extraction process includes three steps. First the Flip Invariant Sketch Histogram (FISH) descriptor is extracted on the input image, which is a flip-invariant version of the shape context feature. Then we explore the symmetry character of the image by calculating the kurtosis coefficient. Finally, the SYM-FISH is generated by constructing a symmetry table. The new SYM-FISH descriptor supplements the original shape context by encoding the symmetric information, which is a pervasive characteristic of natural scene and objects. We evaluate the efficacy of the novel descriptor in two applications, i.e., sketch retrieval and sketch classification. Extensive experiments on three datasets well demonstrate the effectiveness and robustness of the proposed SYM-FISH descriptor.

In this paper, we study the robust subspace clustering problem, which aims to cluster the given possibly noisy data points into their underlying subspaces. A large pool of previous subspace clustering methods focus on the graph construction by different regularization of the representation coefficient. We instead focus on the robustness of the model to non-Gaussian noises. We propose a new robust clustering method by using the correntropy induced metric, which is robust for handling the non-Gaussian and impulsive noises. Also we further extend the method for handling the data with outlier rows/features. The multiplicative form of half-quadratic optimization is used to optimize the non-convex correntropy objective function of the proposed models. Extensive experiments on face datasets well demonstrate that the proposed methods are more robust to corruptions and occlusions.

This paper aims at a newly raising task in visual surveillance: re-identifying people at a distance by matching body information, given several reference examples. Most of existing works solve this task by matching a reference template with the target individual, but often suffer from large human appearance variability (e.g. different poses/views, illumination) and high false positives in matching caused by conjunctions, occlusions or surrounding clutters. Addressing these problems, we construct a simple yet expressive template from a few reference images of a certain individual, which represents the body as an articulated assembly of compositional and alternative parts, and propose an effective matching algorithm with cluster sampling. This algorithm is designed within a candidacy graph whose vertices are matching candidates (i.e. a pair of source and target body parts), and iterates in two steps for convergence. (i) It generates possible partial matches based on compatible and competitive relations among body parts. (ii) It confirms the partial matches to generate a new matching solution, which is accepted by the Markov Chain Monte Carlo (MCMC) mechanism. In the experiments, we demonstrate the superior performance of our approach on three public databases compared to existing methods.

Learning-to-rank for information retrieval has gained increasing interest in recent years. Inspired by the success of sparse models, we consider the problem of sparse learning-to-rank, where the learned ranking models are constrained to be with only a few nonzero coefficients. We begin by formulating the sparse learning-to-rank problem as a convex optimization problem with a sparse-inducing ℓ1constraint. Since the ℓ1constraint is nondifferentiable, the critical issue arising here is how to efficiently solve the optimization problem. To address this issue, we propose a learning algorithm from the primal dual perspective. Furthermore, we prove that, after at most O(1/ε) iterations, the proposed algorithm can guarantee the obtainment of an ε-accurate solution. This convergence rate is better than that of the popular subgradient descent algorithm. i.e., O(1/ε2). Empirical evaluation on several public benchmark data sets demonstrates the effectiveness of the proposed algorithm: 1) Compared to the methods that learn dense models, learning a ranking model with sparsity constraints significantly improves the ranking accuracies. 2) Compared to other methods for sparse learning-to-rank, the proposed algorithm tends to obtain sparser models and has superior performance gain on both ranking accuracies and training time. 3) Compared to several state-of-the-art algorithms, the ranking accuracies of the proposed algorithm are very competitive and stable.

We present an interactive video stylization system for transforming an input video into a painterly animation. The system consists of two phases: a content extraction phase to obtain semantic objects, i.e., recognized content, in a video and establish dense feature correspondences, and a painterly rendering phase to select, place, and propagate brush strokes for stylized animations based on the semantic content and object motions derived from the first phase. Compared with the previous work, the proposed method has the following three advantages. First, we propose a two-pass rendering strategy and brush strokes with mixed colors in order to render expressive visual effects. Second, the brush strokes are warped according to global object deformations, so that the strokes appear to be naturally attached to the object surfaces. Third, we propose a deferred rendering and backward completion method to draw brush strokes on emerging regions and simulate a damped system to reduce stroke scintillation effect. Moreover, we discuss the graphics processing unit-based implementation of our system, which is demonstrated to greatly improve the efficiency of producing stylized videos. In experiments, we verify this system by applying it to a number of video clips to produce expressive oil-painting animations and compare it with the state-of-the-art approaches.

Video shots are often treated as the basic elements for retrieving information from videos. In recent years, video shot categorization has received increasing attention, but most of the methods involve a procedure of supervised learning, i.e., training a multi-class predictor (classifier) on the labeled data. In this paper, we study a general framework to unsupervisedly discover video shot categories. The contributions are three-fold in feature, representation, and inference: (1) A new feature is proposed to capture local information in videos, defined with small video patches (e.g.,$11 \times 11 \times 5$pixels). A dictionary of video words can be thus clustered off-line, characterizing both appearance and motion dynamics. (2) We pose the problem of categorization as an automated graph partition task, in that each graph vertex represents a video shot, and a partitioned sub-graph consisting of connected graph vertices represents a clustered category. The model of each video shot category can be analytically calculated by a projection pursuit type of learning process. (3) An MCMC-based cluster sampling algorithm, namely Swendsen-Wang cuts, is adopted to efficiently solve the graph partition. Unlike traditional graph partition techniques, this algorithm is able to explore the nearly global optimal solution and eliminate the need for good initialization. We apply our method on a wide variety of 1600 video shots collected from Internet as well as a subset of TRECVID 2010 data, and two benchmark metrics, i.e., Purity and Conditional Entropy, are adopted for evaluating performance. The experimental results demonstrate superior performance of our method over other popular state-of-the-art methods.